home *** CD-ROM | disk | FTP | other *** search
-
- ΓòÉΓòÉΓòÉ 1. Version Notice ΓòÉΓòÉΓòÉ
-
- First Edition, September 1993
-
- This edition applies to Version 1 of:
-
- SearchManager/2 (5622-018)
- SearchManager LAN Server/2 and Client/2 (5621-404)
- and to all subsequent releases and modifications until otherwise indicated in
- new editions or technical newsletters.
-
- Order publications through your IBM representative or the IBM branch office
- serving your locality. Publications are not stocked at the address below.
-
- IBM welcomes your comments. A form for readers' comments may be provided at the
- back of this publication, or you may address your comments to the following
- address:
-
- IBM Deutschland Entwicklung GmbH
- Information Development, Dept. 0446
- Postfach 1380
- 71003 Boeblingen
- Germany
-
- When you send information to IBM, you grant IBM a nonexclusive right to use or
- distribute the information in any way it believes appropriate without incurring
- any obligation to you.
-
-
- ΓòÉΓòÉΓòÉ 2. Notices ΓòÉΓòÉΓòÉ
-
- References in this publication to IBM products, programs, or services do not
- imply that IBM intends to make these available in all countries in which IBM
- operates. Any reference to an IBM product, program, or service is not intended
- to state or imply that only IBM's product, program, or service may be used. Any
- functionally equivalent product, program, or service that does not infringe any
- of IBM's intellectual property rights may be used instead of the IBM product,
- program, or service. Evaluation and verification of operation in conjunction
- with other products, except those expressly designated by IBM, is the user's
- responsibility.
-
- IBM may have patents or pending patent applications covering subject matter in
- this document. The furnishing of this document does not give you any license to
- these patents. You can send license inquiries, in writing, to the IBM Director
- of Commercial Relations, IBM Corporation, Purchase, NY 10577.
-
-
- ΓòÉΓòÉΓòÉ 2.1. Trademarks and service marks ΓòÉΓòÉΓòÉ
-
- The following terms, denoted by an asterisk (*) in this publication, are
- trademarks of the IBM Corporation in the United States or other countries or
- both:
-
- BookManager
- DB2/2
- IBM
- Operating System/2
- OS/2
- Personal System/2
- SearchManager/2
- SearchManager LAN Server/2 and Client/2
- SearchManager Client/DOS for Windows
- Thesaurus End User System Toolkit/2.
-
- The following terms, denoted by a double asterisk (**) in this publication, are
- trademarks of other companies:
-
- Ami Pro Samna Corporation
- Intel Intel Corporation
- Lotus Notes Lotus Development Corporation
- Microsoft Microsoft Corporation
- Microsoft Word Microsoft Corporation
- Novell Novell Inc.
- NetWare Novell Inc.
- Windows Microsoft Corporation
- WordPerfect WordPerfect Corporation.
-
-
- ΓòÉΓòÉΓòÉ 3. Read this first ΓòÉΓòÉΓòÉ
-
- This book introduces the IBM* licensed programs:
-
- o IBM SearchManager LAN Server/2 and Client/2* (5621-404)
-
- o IBM SearchManager/2* (5621-018).
-
- SearchManager LAN Server/2 and Client/2 is the client/server version of
- SearchManager/2, for installation in a local area network (LAN).
- SearchManager/2 is the single-workstation version [IBM licensed program IBM
- SearchManager Client/DOS for Windows* (5622-079) is also available] .
-
- These SearchManager/2 products are information retrieval programs for the IBM
- Operating System/2* (OS/2*). They let you search text documents without
- requiring them to be structured in a particular way. Wherever there is need for
- direct access to specific items in a large number of documents such as
- articles, reports, contracts, laws, general directives, or abstracts of any
- type of publication, SearchManager/2 can help you find and extract the relevant
- documents. This applies to documents that are stored in the OS/2 file system,
- or in a database, or in a library.
-
-
- ΓòÉΓòÉΓòÉ 3.1. About this book ΓòÉΓòÉΓòÉ
-
- You do not need previous knowledge of information-retrieval programs to work
- with this book. Information is given for both novice and expert users.
-
- To install the SearchManager/2 products, read Installation.
-
- Before you can search for documents, they need to be indexed. This book is
- written for two audiences: those who are interested in using SearchManager/2 to
- search for information, and those who are concerned with indexing the
- documents.
-
- To search for information, read:
-
- o Introducing IBM SearchManager/2.
-
- o Getting started, skipping the section that concerns indexing.
-
- o Searching for documents.
-
- o Working with found documents.
-
- To index documents, read:
-
- o Introducing IBM SearchManager/2.
-
- o Getting started.
-
- o Planning your document index.
-
- o Putting documents in the queue for indexing.
-
- o Administration.
-
- The linguistic feature of SearchManager/2 is one of the points that make it an
- advanced product. SearchManager/2 uses this feature to make an "intelligent"
- search for documents containing particular terms, and also in creating an
- efficient document index. You can find information about this topic in
- Linguistic processing.
-
- Thesaurus End User System Toolkit/2* (5622-081) is a separately available
- product that you can install together with SearchManager/2. It provides access
- to a thesaurus so that you can enrich your queries with additional search
- terms. This is described in Thesaurus End User System Toolkit/2.
-
- Messages displayed by the SM2 command are explained in Messages displayed by
- the SM2 command.
-
- If a problem should arise, you can find some useful hints and tips in
- Troubleshooting.
-
-
- ΓòÉΓòÉΓòÉ 3.2. Getting to know OS/2 ΓòÉΓòÉΓòÉ
-
- Before you start working with SearchManager/2 you should be familiar with OS/2.
- If you are not, refer to the OS/2 Start here object, and the introductory OS/2
- Tutorial provided with the OS/2 Information object.
-
- You should be familiar with terms such as object, folder, pop-up menu,
- pull-down menu, and window, and with working with a mouse.
-
-
- ΓòÉΓòÉΓòÉ 3.3. Optional IBM licensed programs ΓòÉΓòÉΓòÉ
-
- o IBM SearchManager Client/DOS for Windows (5622-079). SearchManager LAN
- Server/2 and Client/2 is intended for a client/server system in which the
- client and the server workstations are using the OS/2 operating systems. IBM
- SearchManager Client/DOS for Windows is for installation on a DOS client
- workstation to enable it to work with SearchManager LAN Server/2 installed on
- an OS/2 server workstation.
-
- o IBM Thesaurus End User System Toolkit/2 (5622-081) to use a thesaurus when
- you make a query.
-
- o IBM TCP/IP Version 1.2.1 for OS/2 (5798-RXW). If you are installing IBM
- SearchManager LAN Server/2 and Client/2, you can use this product for
- client/server communication, instead of OS/2 Named Pipes.
-
- o IBM DataBase 2 for OS/2 (DB2/2*) (5622-044). You can use this product to
- process the Library Services sample described in the manual IBM
- SearchManager/2: Programming Interfaces.
-
- o Language Dictionaries from the IBM Desktop Dictionary Series.
-
- These dictionaries are available in European, Middle East, and African
- countries, order number 5604-356:
-
- Danish Norwegian
- Dutch Portuguese
- French (National) Spanish
- German Swedish
- Swiss German UK English
- Icelandic US English
- Italian
-
- These dictionaries are available in USA, order number 5799-QBQ:
-
- Danish Italian
- Dutch Norwegian
- French (National) Portuguese
- French (Canadian) Spanish
- German Swedish
-
- Contact your IBM Branch Office or dealer to get information about the
- availability of the dictionaries in your country. Refer to Deciding which
- dictionaries to install.
-
-
- ΓòÉΓòÉΓòÉ 4. Introducing IBM SearchManager/2 ΓòÉΓòÉΓòÉ
-
- Increasing volumes of information are being handled each year, making it more
- and more expensive and time-consuming to file and retrieve documents manually.
- SearchManager/2 is designed to retrieve information that may be spread over
- many and different kinds of documents like letters, notes, newspaper articles,
- reports, contracts, and law books. In the following, the term "documents" is
- used in this broad sense.
-
- The information that you need to find may have been stored in personal
- documents, or in documents that are shared by a department or work group, or
- even in enterprise-wide documents.
-
-
- ΓòÉΓòÉΓòÉ 4.1. If you are new to retrieval ΓòÉΓòÉΓòÉ
-
- Imagine that you work for a travel agency; the agency's contracts and
- correspondence with customers are prepared using a word-processing program,
- such as WordPerfect**, or Ami Pro**; internal correspondence is written in flat
- ASCII files.
-
- Then assume that you need to search through these electronic documents to find
- those that concern price increases.
-
- Using SearchManager/2, you specify a query consisting of two search terms price
- and increase, linked by AND to indicate that you want to find documents that
- contain both of these terms.
-
- SearchManager/2 searches and displays the names of the documents that match
- your search criteria. You can then display any of these documents; the terms
- that you searched for are highlighted.
-
- This is retrieval in its simplest form; but it applies to environments ranging
- from, say, a freelance translator working at a single workstation, using only
- one document type, to a large company whose employees use workstations
- connected to a complex local area network, and whose documents are in many
- different formats and are stored at many different workstations.
-
- Using a query to search for text in many documents
-
- SearchManager /
- 2canrapidlyretrieverelevantinformationfromlargecollectionsofdocuments
- ,usingcombinationsofwordsassearchterms .
-
- Unlike database-management systems, SearchManager/2 does not require that data
- be rigorously structured into interrelated files, records, and fields, before
- it can be accessed. Instead, SearchManager/2 operates by building a
- comprehensive index of significant words or terms that appear in a collection
- of documents. Any combination of these terms can then be used to find the
- documents that contain them.
-
- SearchManager/2 features high-performance indexing and retrieval. It can access
- word-processing documents in their original native form, and offers a rich set
- of retrieval capabilities including word, phrase, "wild-card", and proximity
- searching using full Boolean logic.
-
-
- ΓòÉΓòÉΓòÉ 4.2. Local area network or single workstation? ΓòÉΓòÉΓòÉ
-
- There are two SearchManager/2 products: IBM SearchManager LAN Server/2 and
- Client/2, and IBM SearchManager/2. SearchManager LAN Server/2 and Client/2 is
- intended for use in local area networks.
-
- Finding documents using SearchManager LAN Server/2 and Client/2
-
- There is also a need to find information stored on a personal "stand-alone"
- workstation. This is the area for which the SearchManager/2 single-workstation
- version is designed.
-
- Finding documents using SearchManager/2 single workstation
-
-
- ΓòÉΓòÉΓòÉ 5. Getting started ΓòÉΓòÉΓòÉ
-
- If you are in a hurry to get started, follow the steps in this chapter; it will
- help you to become familiar with SearchManager/2 quickly. More detailed
- information about each step is given elsewhere in this book.
-
- After a brief introduction to the help system, this chapter takes you through
- indexing some sample documents and then through a simple search scenario, where
- you locate and then display a found document.
-
- Tip If you want to get to the section on searching for documents right
- away, turn to Making your first search. Be sure, though, that documents
- have already been indexed.
-
-
- ΓòÉΓòÉΓòÉ 5.1. Getting help when you need it ΓòÉΓòÉΓòÉ
-
- SearchManager/2 has an extensive Help facility.
-
- The information line in each window: At the bottom of each window, permanently
- displayed, there is a one-line information area that tells you about the field
- at the current cursor position.
-
- Getting more help for a field: If you need more information than is shown in
- the information area:
-
- o Select the field for which you need information
- o Press F1, or select the Help push button.
-
- A help window is displayed giving you detailed information about the field at
- the current cursor position.
-
- Getting help for a window: To display help information about a particular
- window, rather than about a particular field, from any Help window you can
- press F2.
-
- Other ways of getting help: Whenever you are working with a window that has a
- menu bar, select the Help choice to display the various types of help you can
- get. Simply click on each of them in turn to find out what each of them
- offers.
-
- If the window does not have a menu bar, press the Help push button, or select
- Help from the system-menu symbol - the icon in the top left-hand corner of the
- window.
-
- Getting more details about messages: While you are working with
- SearchManager/2, information messages are occasionally displayed. If you need
- more information about the message, select Help on the message window.
-
- Moving easily through the help system: Help windows contain words that are
- highlighted in a different color. If you double-click on any of these words, a
- new help window is displayed that provides information about the highlighted
- words.
-
- Viewing the online documentation:
-
- Tip If this online book covers important parts of your screen, resize its
- window. The text is automatically reflowed within the new window size.
-
-
- ΓòÉΓòÉΓòÉ 5.2. A brief look at the indexing process ΓòÉΓòÉΓòÉ
-
- In this section, you:
-
- 1. Open the SearchManager/2 folder
-
- 2. Place the sample documents in the queue for indexing
-
- 3. Index the sample documents.
-
- If you are working with a client/server system, the sample documents are
- installed at the server workstation. You must have access to these documents.
-
-
- ΓòÉΓòÉΓòÉ 5.2.1. Open the SearchManager/2 folder ΓòÉΓòÉΓòÉ
-
- The SearchManager/2 folder is an object on the OS/2 Workplace Shell desktop.
-
- 1. Double-click on the SearchManager/2 object.
-
- Figure: The SearchManager/2 folder
-
- The SearchManager/2 folder contains the following objects:
-
- o A Search service object; this does the indexing and retrieval.
-
- o A Retrieval and Indexing object; this is an online version of this manual.
-
- o An Installation Utility object to assist you in installing components of
- SearchManager/2.
-
-
- ΓòÉΓòÉΓòÉ 5.2.2. Place the sample documents in the queue for indexing ΓòÉΓòÉΓòÉ
-
- The names of any documents that you want to have indexed must be placed in the
- document queue. To do this:
-
- 1. Open the Search service folder to display the Document index object, the
- Sample documents folder, and a Query object.
-
- Figure: The Search service folder
-
- Sampledocumentsareprovidedfortestpurposes ;theyareinaSampledocumentsfolder
- .( Inaclient / serversystem
- ,thesampledocumentsareinstalledonlyattheserverworkstation . )
-
- 2. Use the mouse to drag the Sample documents folder and drop it onto the
- document index object. The Add to Queue window opens.
-
- 3. Select Queue in the Add to Queue window.
-
- The names of the sample documents are now in a queue, ready to be indexed. A
- queue is used because indexing uses a lot of processing power, and is
- typically done at a time when the workstation is not very busy.
-
-
- ΓòÉΓòÉΓòÉ 5.2.3. Index the sample documents ΓòÉΓòÉΓòÉ
-
- (If you are working with a client/server system, the system administrator does
- this step.)
-
- To confirm that the document names are in the queue:
-
- 1. On the index object, click the right-hand mouse button to display its
- pop-up menu.
-
- 2. Select Administration.
-
- The Administration notebook shows how many documents there are in the queue.
- To index these:
-
- 1. Click on the Index push button.
-
- 2. Click on OK in the Confirmation window.
-
- 3. After half a minute or so, click on the Refresh push button to see if
- indexing has finished. Repeat this until the number of documents in the
- queue becomes zero.
-
- The Administration notebook now indicates how many documents there are in the
- index. The first time you index any documents, the indexed terms are placed in
- the Primary part of the document index. Therafter, the Secondary part is used.
-
- o Close the Administration notebook (select Close from the system-menu symbol,
- or double-click on it).
-
-
- ΓòÉΓòÉΓòÉ 5.3. Making your first search ΓòÉΓòÉΓòÉ
-
- This section assumes that the previous step, indexing documents, has already
- been done.
-
- In this section, you:
-
- o Use a query object to search for documents
- o Browse one of the found documents
- o Close the query.
-
-
- ΓòÉΓòÉΓòÉ 5.3.1. Search for documents ΓòÉΓòÉΓòÉ
-
- Figure: A query object contains the query and the results
-
- Here you search for documents containing the terms cat and mouse:
-
- 1. Open the Query object provided in the Search service folder. The Query
- View notebook is displayed.
-
- 2. In the first Search terms field, type cat.
-
- 3. In the Operator field, click on the drop-down button to display the list
- of operators.
-
- 4. Select the AND operator.
-
- 5. In the second Search terms field, type mouse.
-
- 6. Select the Search push button to run the query.
-
- Figure: The Query View notebook
-
- The documents that contain both of these search terms are shown in the Results
- View window. If your document index contains only the sample documents, then
- only one document is found.
-
-
- ΓòÉΓòÉΓòÉ 5.3.2. Browse the found document ΓòÉΓòÉΓòÉ
-
- 1. Double-click on the found document object to open it and to display the
- text of the document.
-
- 2. In the Browser window select Next match to find the first occurrence of
- the search terms. The search terms are highlighted.
-
- Figure: Browsing a document
-
-
- ΓòÉΓòÉΓòÉ 5.3.3. Close the query ΓòÉΓòÉΓòÉ
-
- 1. Close the Browser window.
-
- 2. Close the Results View of the query object.
-
- 3. Close the Query View of the query object.
-
- The two views of a query object, the Query View and the Results View, are
- available from the object's pop-up menu as shown in the figure in Search for
- documents. From here, you can also copy the query or create another.
-
-
- ΓòÉΓòÉΓòÉ 6. Planning your document index ΓòÉΓòÉΓòÉ
-
- This chapter helps you to plan before you index documents for the first time.
- It explains:
-
- o Why documents need to be indexed
- o Which document types you can index
- o Which document libraries and file systems you can use
- o How to decide which type of index to use.
-
-
- ΓòÉΓòÉΓòÉ 6.1. Why documents need to be indexed ΓòÉΓòÉΓòÉ
-
- A fast information retrieval system does not sequentially scan through
- documents; this would take too long. Instead, it operates on a previously built
- document index. You can think of a document index as consisting of terms
- extracted from the documents, stored together with the document names.
-
- The retrieval system searches through the index for the terms requested, and
- finds the names of the documents containing those terms.
-
- Figure: Rapid retrieval by indexing only meaningful terms
-
- Before starting to search for and retrieve information, you must create an
- index from the documents to be processed. Because indexing can be a lengthy
- process, it is done in two steps:
-
- 1. First, you place the names of the documents to be indexed in the document
- queue.
-
- 2. Then, usually when the workstation is not busy with other activites, you
- index the documents listed in the queue.
-
- Figure: Indexing is a two-step process
-
- A document index contains only relevant information; insignificant information
- (stop words), like conjunctions and prepositions are not indexed.
- SearchManager/2 uses a list of these stop words to prevent them from being
- indexed.
-
-
- ΓòÉΓòÉΓòÉ 6.2. Document types that can be indexed ΓòÉΓòÉΓòÉ
-
- Here is a list of the types of documents that you can index:
-
- o Flat ASCII
- o Ami Pro Architecture Version 4
- o IBM Final Form Text (FFT): Document Content Architecture
- o IBM Revisable Form Text (RFT): Document Content Architecture
- o IBM SearchManager/2 Text format (for more information, see IBM
- SearchManager/2: Programming Interfaces)
- o Microsoft** Rich Text Format (RTF), Version 1
- o WordPerfect (OS/2 and Windows**), Versions 5.0, 5.1, and 5.2
- o Microsoft Word**, Versions 5.0 and 5.5
- o Microsoft Word for Windows, Versions 1.0, 1.1, and 2.0
- o IBM SearchManager/370 Editable Condensed Text Format (ECTF); each document
- must be specified in CCSID 4946 (code page 850) as a separate OS/2 file.
-
- When indexing a document from the OS/2 file system, SearchManager/2 can detect
- its type automatically. But SearchManager/2 also needs to know what language
- the document is written in, and its Coded Character Set Identifier (CCSID).
- For this, SearchManager/2 uses information that you specify in the Rules
- section of the index Administration notebook. See Defining indexing rules for
- OS/2 document types.
-
- When SearchManager/2 indexes documents from a Lotus Notes** library system,
- these can be in plain text or in RTF format. If they are in RTF format,
- SearchManager/2 extracts plain text from them.
-
-
- ΓòÉΓòÉΓòÉ 6.3. Libraries and file systems ΓòÉΓòÉΓòÉ
-
- Using SearchManager/2, you can index and search for documents in:
-
- o The OS/2 file system
-
- o Lotus Notes Version 3
-
- o Other text libraries or applications, using the user exit library services
- described in IBM SearchManager/2: Programming Interfaces.
-
- Note: The names of documents or document groups differ depending on the
- library used. When using a library other than the OS/2 file system, only
- Browse and Add to queue are available for documents contained in a result.
-
- o Databases, such as the IBM Database 2 for OS/2, supported by the library
- services sample.
-
- The examples in this book refer to the OS/2 file system.
-
- Tip For each file system or library system, a separate SearchManager LAN
- Server/2 installation is required. Each has to be installed on a separate
- server.
-
-
- ΓòÉΓòÉΓòÉ 6.4. Deciding which type of document index to use ΓòÉΓòÉΓòÉ
-
- You can choose one of three index types:
-
- o Linguistic
- o Precise
- o Dual.
-
- This section summarizes the index types. If you need a more detailed
- description of how each type affects linguistic processing in SearchManager/2,
- read Linguistic processing. Specifying properties for the document index
- describes how you can specify which index type you want to have.
-
- Tip Choosing an index type is a critical decision for an administrator;
- changing to or from a linguistic index type involves clearing the index
- and indexing all documents again.
-
-
- ΓòÉΓòÉΓòÉ 6.4.1. Linguistic index ΓòÉΓòÉΓòÉ
-
- If you choose to create a linguistic index, then, during indexing, linguistic
- processing is applied while analyzing the documents' text.
-
- For a query, SearchManager/2 then applies the same linguistic processing to the
- search terms before searching in the document index. The result is that any
- form of a search term matches any other form occurring in one of the indexed
- documents.
-
- For example, the search term mouse matches the document terms mouse, mice, MICE
- (capital letters), and so on. Similarly, the search term Mice matches the same
- document terms.
-
- The advantage of this type of index is that more documents are likely to be
- found (increased "recall").
-
- This index type requires the least amount of disk space. However, indexing and
- searching may take longer than for the precise index.
-
- The linguistic processing used to index documents for a linguistic index are:
-
- o Word and sentence separation.
-
- o Changing terms to a standard form, in which there are no capital letters,
- and accented letters like Б are changed to a form without accents
- ("normalization"). For example, GrБсe is indexed as gruesse.
-
- o Reducing terms to their base form ("lemmatization"). For example, bought is
- indexed as buy, mice as mouse.
-
- o Word decomposition, where compound words like the German Wetterbericht
- (weather report) are indexed not only as wetterbericht, but also as wetter
- and bericht.
-
- o Stop-word filtering in which only the relevant terms are extracted for
- indexing. A report about all animals is indexed as report and animal.
-
- o Part-of-speech filtering, which is similar to stop-word filtering; only
- nouns, verbs, and adjectives are indexed. I drive my car quickly is indexed
- as drive and car. The words I and my are removed as stop words, but
- additionally the word quickly is removed.
-
- The first two of these linguistic processes, that is, word and sentence
- separation, and normalization, are possible without having dictionaries
- installed.
-
- Tip Even word fragments (words masked with global characters) are compared
- to the base forms in the index. For example, a document containing the
- text
-
- I swum
-
- being searched for all words beginning with
-
- swu
-
- using the word fragment
-
- swu*
-
- is not found, because the index contains only the term
-
- swim
-
- So the query should be
-
- swi*
-
- instead.
-
- The table in Linguistic processing for indexing summarizes how terms are
- extracted for indexing when you are using a linguistic index.
-
-
- ΓòÉΓòÉΓòÉ 6.4.2. Precise index ΓòÉΓòÉΓòÉ
-
- In a precise index, linguistic processes are used only to determine word and
- sentence boundaries. The terms in the documents are indexed in exactly the same
- form as they occur in the document.
-
- For example, the search term mouse matches only the document term mouse and
- not, for example, MICE.
-
- The advantage of this type of index is that the search is more precise (reduced
- recall); also indexing and retrieval is faster. Since each different form and
- spelling of every term is indexed, more disk space is needed.
-
- The linguistic processes used to index documents for a precise index are:
-
- o Word and sentence separation
- o Sentence-begin processing
- o Stop-word filtering.
-
- These are explained in Linguistic index,and you can find more details in
- Linguistic processing. Linguistic processing for indexing gives some examples
- of how terms are extracted from document text for indexing when you are using
- a precise index.
-
- In a query, the same processing is applied to the query terms, which are then
- compared with the terms found in the index.
-
- This means that only the same form or inflection of a search term is found.
- Generally, the search in a precise index is case sensitive.
-
-
- ΓòÉΓòÉΓòÉ 6.4.3. Dual index ΓòÉΓòÉΓòÉ
-
- The dual index is a precise index with linguistic support. This is the most
- flexible index type, because it allows you to decide for each search term
- whether to search linguistically or precisely.
-
- This index type needs the most disk space, however, and indexing and searching
- are slower than for a linguistic index.
-
- A dual index contains the normalized form (standard form in lowercase letters,
- and without accents), the lemmatized form (base form, such as the infinitives
- of verbs), and the original form of each term.
-
- In a query, you can choose the processing that is applied to the query terms:
-
- o Linguistic search (Inflected forms option selected); any form and any case
- matches any other form and any other case.
-
- o Precise search (Inflected forms not selected); only the exact form and exact
- case matches are searched for.
- In a dual index, word fragments are always looked for precisely; matching for
- word fragments is case sensitive.
-
-
- ΓòÉΓòÉΓòÉ 7. Putting documents in the queue for indexing ΓòÉΓòÉΓòÉ
-
- Indexing is a two-step process. The first step is to put the names of the
- documents to be indexed into a queue; this is covered in this chapter. The
- second step, often done later, is to index those documents named in the queue.
-
- Removing terms from an index follows the same two-step process. You would do
- this when, for example, you intend to delete documents from your library.
-
- The concept of indexing is described in Why documents need to be indexed, and
- how to do it is described in Administration.
-
- You can place individual documents or document groups (folders) in the document
- queue for indexing. There are several ways to do this:
-
- o By double-clicking on the index object
- o By dragging folders or individual documents to the index object
- o By using the pop-up menu of documents in a result
- o Automatically, by having documents indexed periodically.
-
- In each case, you can choose whether to have the documents' terms added to the
- index (Add), or, if the documents have already been indexed, to have the
- documents' terms removed from the index (Remove).
-
- The final section in this chapter describes queuing Lotus Notes databases.
-
- Tip If you are using a client/server system, be sure that the client and
- the server workstations have access to documents marked for indexing.
-
-
- ΓòÉΓòÉΓòÉ 7.1. Queuing by double-clicking on the index object ΓòÉΓòÉΓòÉ
-
- The first method of queuing documents is to double-click on the Index object to
- display the Add to queue window.
-
- Then specify what is to be placed in the document queue by selecting the
- document groups and their documents displayed on the Add to Queue window.
- Double-click on a document group to display its subgroups.
-
- Figure: The Add to Queue window
-
-
- ΓòÉΓòÉΓòÉ 7.2. Queuing by dragging folders or documents to the index object ΓòÉΓòÉΓòÉ
-
- An alternative method of queuing documents is to select documents or document
- groups, then use the mouse to drag them and drop them onto the Index object.
-
- The selected documents and the documents contained in the selected document
- groups are then displayed in a document list on the Add to Queue window for
- further processing.
-
-
- ΓòÉΓòÉΓòÉ 7.3. Queuing using the pop-up menu of documents in a result ΓòÉΓòÉΓòÉ
-
- On occasions, you may decide that a document that you have found by a search
- needs to be indexed again, either because the indexed version is out of date,
- or because you want the document's terms to be removed from the index.
-
- To place this document in the queue for indexing:
-
- o In the Results View, select the Add to queue choice in the pop-up menu of
- the Document (described in Working with documents in the Result View).
-
-
- ΓòÉΓòÉΓòÉ 7.4. Automatic queuing using periodic indexing ΓòÉΓòÉΓòÉ
-
- The methods described so far are useful when you want to index particular
- documents at a particular time. But periodic indexing is likely to be the
- method you will use most.
-
- This method lets you forget about indexing; it happens automatically. Typically
- the documents in which you want to search are located in document groups or
- folders, and these documents are continually changing or being added to.
-
- Using periodic indexing, you simply specify which document groups you want to
- have indexed, and at which intervals. At the required time, SearchManager/2
- automatically places these document groups in the queue for indexing and
- immediately afterwards indexes them. How to do this is described in Indexing
- documents periodically.
-
- Tip If you intend to index periodically, you should do it from the
- beginning, because the first time you index periodically, all the
- documents in a document group are indexed, even if they have been indexed
- before.
-
-
- ΓòÉΓòÉΓòÉ 7.5. Queuing Lotus Notes databases ΓòÉΓòÉΓòÉ
-
- To put Lotus Notes databases in the document queue for indexing, use the index
- object as described in Queuing by double-clicking on the index object. In the
- Add to Queue window, the Groups list shows servers, subdirectories, and
- databases; the Documents list shows a list of symbolic document names
- corresponding to the Lotus Notes naming convention.
-
- Lotus Notes documents are typically small, so it is not reasonable to index
- them individually; select all the documents listed.
-
- You can also have these databases placed in the queue automatically. Indexing
- documents periodically describes how to do this.
-
-
- ΓòÉΓòÉΓòÉ 8. Administration ΓòÉΓòÉΓòÉ
-
- Before you read this chapter, you should be familiar with Why documents need to
- be indexed. This chapter describes how you can speed up indexing and retrieval;
- it introduces the two parts of a document index and the concepts of Merge and
- Cleanup.
-
- This chapter then covers the following index administration activities:
-
- o Tailoring the document index: telling SearchManager/2 how to distinguish
- among various document types, specifying which stop-word list to use, and
- which type of index this is.
-
- o Document index administration: indexing, merging, and cleaning up the index.
- These last two activities make indexing and retrieval faster, and reduce the
- amount of storage space needed by the index.
-
- o Periodic document index administration: indexing, merging, and cleaning up
- the index automatically at intervals.
-
- o Displaying and changing the status of SearchManager/2 activities.
-
- Tip If you are working with SearchManager LAN Server/2 and Client/2, only
- the LAN administrator can work with the document index; you can do this at
- either a client or the server workstation.
-
- The three index-administration activities that you do most are:
-
- Indexing documents
- Merging the two parts of the document index
- Cleaning up the document index.
-
- Tip Searching can still continue while these administration activities are
- being carried out. Administration is mostly concerned with maintaining the
- document index. In a client/server system, however, the administrator is also
- responsible for starting and stopping the search service, the part of
- SearchManager/2 that handles indexing and retrieval.
-
-
- ΓòÉΓòÉΓòÉ 8.1. Starting and stopping a search service ΓòÉΓòÉΓòÉ
-
- The SearchManager/2 search service usually starts automatically when you search
- for documents or start any administration task. For this reason, your first
- query, for example, may take a little longer than usual. However, if you are
- working with a client/server system, the system administrator must start the
- SearchManager/2 search service at the server workstation so that it can be used
- by the client workstations.
-
- Tip If you do administration periodically, then, to be sure that the
- search service is always started, start it automatically using either the
- STARTUP.CMD method, or the Startup folder method.
-
- You start a search service, at the workstation on which it is installed, in
- one of the following ways:
-
- o Using the Search service folder. To start the search service immediately,
- select Start on the pop-up menu of the Search service folder.
-
- o Using the SM2 command. Start the search service in an OS/2 session by
- entering the command
-
- sm2 start servicename
-
- at the OS/2 prompt.
-
- servicename is the name specified in the Definitions page of the search
- service object.
-
- o Using the STARTUP.CMD file. To start the search service automatically at
- system startup, place the statement
-
- sm2 start servicename
-
- in the STARTUP.CMD batch file.
-
- o Using the OS/2 Startup folder. An alternative way to start the search
- service automatically at system startup is to create a program object for
- SM2.EXE, having the parameters
-
- start and servicename
-
- Then place this object in the Startup folder of your OS/2 system.
-
- Similarly, you can stop a search service in one of the following ways:
-
- o Using the Search service folder. Select Stop on the pop-up menu of the
- Search service folder
-
- o Using the SM2 command. In an OS/2 session, enter the command
-
- sm2 stop servicename
-
-
- ΓòÉΓòÉΓòÉ 8.1.1. Displaying the status of a search service ΓòÉΓòÉΓòÉ
-
- To display the status of a search service, whether or not it is running,
- display the pop-up menu of the Search service folder. If the Stop choice is
- available, then the search service is running. If the Start choice is
- available, then the search service is not running.
-
- Alternatively, you can enter the following OS/2 command:
-
- sm2 status servicename
-
-
- ΓòÉΓòÉΓòÉ 8.2. Speeding up indexing and retrieval ΓòÉΓòÉΓòÉ
-
- This section introduces the two parts of a document index, and describes how
- merge and cleanup can keep the index tuned for maximum efficiency, and help you
- to save disk space.
-
-
- ΓòÉΓòÉΓòÉ 8.2.1. The two parts of an index ΓòÉΓòÉΓòÉ
-
- A document index has a primary part and a secondary part; it is designed this
- way to speed up indexing. During a search, SearchManager/2 looks in both parts
- of the index to find a match.
-
- The way you handle these parts can affect how fast your documents are indexed
- and how long a search takes.
-
- Figure: The two parts of a document index
-
-
- ΓòÉΓòÉΓòÉ 8.2.2. Speeding up indexing ΓòÉΓòÉΓòÉ
-
- Each time you index documents, the indexed terms are stored in the secondary
- part of the document index. As the secondary part grows, indexing becomes less
- efficient, so from time to time, you move (merge) the indexed terms from the
- secondary part to the primary part.
-
- After a time, you have the situation shown in The two parts of an index where
- the primary index has grown large and the secondary index remains small.
- Indexing continues to be efficient.
-
- You can merge the secondary part into the primary part at irregular intervals
- when you explicitly request it (see Merging the document index parts), or you
- may find it more convenient to automatically merge the index parts periodically
- (see Merging the index parts periodically). Whichever method you choose, you
- should plan to merge frequently enough to prevent the secondary part of the
- index reaching a size that significantly slows down indexing.
-
- When the primary index is relatively small, always aim to keep the secondary
- part smaller than the primary part. As the primary index grows, you will gain
- experience and will be able to judge for yourself when merging the indexes
- significantly improves the speed of indexing.
-
-
- ΓòÉΓòÉΓòÉ 8.2.3. Speeding up retrieval ΓòÉΓòÉΓòÉ
-
- This section applies only if you are using the OS/2 file system.
-
- Before you delete a document from your library, you can add it to the document
- queue to have its terms removed from the document index.
-
- But removing terms in this way is not always convenient; often it is easier to
- clean up the document index from time to time after the documents have been
- deleted, removing those terms that belong to documents that you have deleted.
- During cleanup SearchManager/2 checks for each term whether the document that
- it belonged to still exists. If the document does not exist, SearchManager/2
- removes the term from the index. (If the search service no longer has access to
- a particular resource, and therefore cannot determine whether documents have
- been deleted, cleanup is not done for those documents.)
-
- Cleanup reduces the size of your index making retrieval faster and, of course,
- saving disk space.
-
- You can clean up the index at irregular intervals when you explicitly request
- it (see Cleaning up the document index), or you can do it automatically (see
- Cleaning up periodically).
-
-
- ΓòÉΓòÉΓòÉ 8.3. Opening the Administration notebook ΓòÉΓòÉΓòÉ
-
- To display the Administration notebook, select Administration in the pop-up
- menu of the Index object.
-
- Figure: The Administration notebook
-
-
- ΓòÉΓòÉΓòÉ 8.4. Tailoring the document index ΓòÉΓòÉΓòÉ
-
- This section tells you how to specify:
-
- o The rules that enable SearchManager/2 to distinguish among various types of
- documents of the OS/2 file system and the language in which they are
- written.
-
- o The stop words used when indexing.
-
- o The properties of the index, including the index type: precise, linguistic,
- or dual.
-
- Tip You should tailor the document index with care before you index any
- documents. Once the document index contains terms, it can be difficult to
- change these tailoring parameters without completely clearing the document
- index first.
-
- To tailor the document index:
-
- 1. Open the search service folder.
-
- 2. On the pop-up menu of the Index object, select Administration.
- This displays the Administration notebook, which has sections for specifying
- the indexing rules, stop words, and properties of the document index.
-
-
- ΓòÉΓòÉΓòÉ 8.4.1. Defining indexing rules for OS/2 document types ΓòÉΓòÉΓòÉ
-
- When indexing an OS/2 document, SearchManager/2 needs to know:
-
- o Which type of document it is; if it is a WordPerfect document, for example,
- or a flat ASCII file. There is a list of document types that SearchManager/2
- can work with in Document types that can be indexed.
-
- o Which language is the document written in.
-
- o Which Coded Character Set Identifier (CCSID) is used.
-
- SearchManager/2 can detect most of the listed document types automatically. If
- for some reason automatic recognition is not possible, then SearchManager/2
- needs to refer to rules (or file-naming conventions) that you establish to
- relate document types to file extensions.
-
- For example, you could assign the extension .SAE to Ami Pro documents written
- in English using the international CCSID 850, and you could assign the
- extension .SAF to Ami Pro documents written in French using the same CCSID.
-
- You establish these rules in the Rules section of the index Administration
- notebook.
-
- Tip These rules are also used to determine which documents are indexed
- during periodic indexing (in addition to other parameters that you
- specify). Establish these rules thoroughly before running periodic
- indexing; this can prevent unwanted documents being indexed.
-
- To define the indexing rules:
-
- o Select the Rules section.
-
- Figure: The Rules section
-
- New rule: Here you can specify:
-
- o Which files are not to be indexed, such as files with EXE or OBJ file
- extensions.
-
- Tip Many document extensions have already been set up. Check that these
- match the extensions used by your documents and make any changes
- necessary.
-
- o Which language, format, and CCSID applies for the specified document type.
- Here is a list of CCSIDs:
-
- 37 USA, Canada, and others
- 273 Austria, Germany
- 277 Denmark, Norway
- 278 Finland, Sweden
- 280 Italy
- 284 Spain, Latin America
- 285 United Kingdom
- 297 France
- 437 USA
- 500 International Latin-1
- 819 Latin-1
- 850 Latin-1
- 860 Portugal
- 861 Iceland
- 863 Canada
- 865 Denmark, Norway
- 871 Iceland
- 4946 Latin-1 (Codepage 850)
-
- Default rule: Which format, language, and CCSID that SearchManager/2 is to
- assume if the document extension is none of those specified. This format is
- used only if SearchManager/2 cannot detect this automatically.
-
- Tip If your documents are all the same type, are all in the same language,
- and all use the same CCSID, specify only a default rule. But consider also
- setting up rules to exclude programs and other files that are not
- documents.
-
-
- ΓòÉΓòÉΓòÉ 8.4.2. Defining indexing rules for Lotus Notes ΓòÉΓòÉΓòÉ
-
- For Lotus Notes documents, define only a default rule:
-
- Format ASCII.
- Language Set according to the installed version of Lotus Notes, or to the
- language in which the documents are written. One language is
- supported.
- CCSID 850, the default for Lotus Notes.
-
-
- ΓòÉΓòÉΓòÉ 8.4.3. Specifying stop words for document indexing ΓòÉΓòÉΓòÉ
-
- Stop words are irrelevant words that you do not want to store in the document
- index. This not only reduces the size of the index and reduces the time it
- takes to search through it, it can also increase the precision of your search.
- For more information, refer to Stop-word filtering.
-
- Each dictionary that you have installed for use with SearchManager/2 provides a
- list of stop words. But you can add stop words to and delete stop words from
- these lists.
-
- If you are using SearchManager LAN Server/2 and Client/2, you store and change
- the stop-word lists while you are working at the workstation on which the
- search service is installed; it cannot be done from a remote client.
-
- To select a stop word list, or to add and remove words in the list:
-
- o Select the Stop words section of the Administration notebook.
-
- o Select Language to choose which dictionary's stop-word list you want to work
- with.
-
- Figure: The Stop words section
-
- Tip Consider extending the list of stop words with common words from your
- business area that are likely to appear in many documents, but are
- unlikely to be searched for. This reduces the size of your index, and
- leads to faster searches.
-
- Here are some points to consider if you add or remove stop words after some
- documents have been indexed:
-
- o When you add stop words, documents that have already been indexed behave
- during a search like those that you index later using the changed stop-word
- list. It is true that any occurrences of the new stop words will have been
- indexed earlier, but the stop-word list now prevents these words from being
- found.
-
- o When you remove stop words, you should consider indexing all your documents
- again. This is because documents that have already been indexed behave
- differently during a search from those that you index later using the
- changed stop-word list. The fact that stop words have been removed means
- that now you can find these words in documents that you index thereafter.
- But these words will not have been indexed for any documents that you
- indexed before this time; so a search does not find them in those documents.
-
-
- ΓòÉΓòÉΓòÉ 8.4.4. Specifying properties for the document index ΓòÉΓòÉΓòÉ
-
- These are the index properties:
-
- o The index type: linguistic, precise, or dual
-
- o The programs to use for the type of library system or file system you are
- using.
-
- To define the document index properties:
-
- o Select the Properties section of the Administration notebook.
-
- Figure: The Properties section
-
-
- Index types
-
- The index types are: linguistic, precise, and dual. These are described in
- Deciding which type of document index to use. Here is a summary:
-
- Linguistic index
- Any inflected form of a search term matches any other inflected form
- occurring in one of the indexed documents. For example, the search
- term mouse matches the document terms mouse, mice, MICE, and so on.
- The search term mice matches the same terms.
- Precise index
- The terms in the documents are indexed in exactly the same form as
- they occur in the document. For example, the search term mouse
- matches only the document term mouse and no other variations, like
- Mouse.
- Dual index
- The dual index is a combination of the other two types. At the time
- of making your search, you can choose whether you want
- SearchManager/2 to search for inflected forms or not.
-
- Changing the document index type: Once you have set the index type you can
- change it, but when changing to or from a linguistic index type, the index
- must first be cleared.
-
- A change from precise to dual or the other way around can be done without
- clearing the index. The change occurs the next time you merge the index parts.
-
- Figure: Changing the index type
-
-
- Library services
-
- In the Library services field, you specify which interface programs are used
- by SearchManager/2 to work with the library or file system you intend to use.
- These programs are in the form of dynamic link libraries (DLLs).
-
- When SearchManager/2 is first installed, these fields contain the names of the
- DLLs for the OS/2 file system:
-
- Client: EHSLSCFS
- Server: EHSLSSFS
-
- SearchManager/2 also provides library service DLLs which support Lotus Notes
- Version 3. If you are using Lotus Notes, enter the following DLL names:
-
- Client: EHSLSCLN
- Server: EHSLSSLN
-
- If you intend to use a different library system, however, you can write your
- own DLLs. For details refer to the manual IBM SearchManager/2: Programming
- Interfaces.
-
- If you decide to change to a different library system, that is, if you change
- the DLLs, clear the document index and then index the new library's documents.
-
- For a server workstation: If you are setting up the library services as part
- of the installation of SearchManager LAN Server/2 on a server workstation, any
- subsequent clients that you install will automatically have the correct
- Library services settings. Of course, if you are using your own library
- services, you must install your client DLLs at the client workstations.
-
- If, however, you have previously installed Client/2 on client workstations, a
- message informs you that the Library service settings have changed, and they
- are automatically updated the next time you use the search service, such as
- when you make a query.
-
-
- ΓòÉΓòÉΓòÉ 8.5. Document index administration ("Explicit") ΓòÉΓòÉΓòÉ
-
- You can index documents either at irregular intervals (that is, when you
- explicitly request it) or at regular intervals (periodically).
-
- To index documents immediately, select the Explicit section of the
- Administration notebook.
-
- Figure: The Explicit section
-
- This section shows the number of documents in the document queue and in the
- primary and secondary parts of the document index. The indicators below the
- number show the size of each relative to the largest.
-
-
- ΓòÉΓòÉΓòÉ 8.5.1. Adding and removing documents in the document index ΓòÉΓòÉΓòÉ
-
- Select Index to process the documents named in the document queue; that is, to
- add their terms to the index, or to remove them, whichever you indicated when
- you placed the documents in the queue.
-
- If the Index push button is unavailable:
-
- o The document queue is already empty
- o Indexing is already running
- o Indexing is locked
- o Merge or cleanup is running.
-
- Check the Status section of the Administration notebook for more information.
-
- Occassionally, while SearchManager/2 is indexing documents, a merge occurs.
- This can happen if the document queue is particularly large; indexing is
- interrupted and a merge takes place, then indexing is resumed.
-
-
- ΓòÉΓòÉΓòÉ 8.5.2. Clearing the document queue ΓòÉΓòÉΓòÉ
-
- Select Clear to empty the document queue. The documents are not indexed.
-
- You will probably not use Clear very often, but it can be useful while you are
- trying out the system after installation.
-
- If the Clear push button is unavailable:
-
- o The document queue is already empty
- o Indexing is running
- o Cleanup is running.
-
- Check the Status section of the Administration notebook for more information.
-
-
- ΓòÉΓòÉΓòÉ 8.5.3. Merging the document index parts ΓòÉΓòÉΓòÉ
-
- Select Merge to merge the indexed terms contained in the secondary part of the
- document index with those in the primary part of the document index. (If you
- are not yet familiar with these terms, turn to Speeding up indexing.) The
- secondary part of the document index is then cleared.
-
- If the Merge push button is unavailable:
-
- o The secondary part of the document index is empty
- o Indexing, cleanup, or merge is running (or locked).
-
- Check the Status section of the Administration notebook for more information.
-
-
- ΓòÉΓòÉΓòÉ 8.5.4. Cleaning up the document index ΓòÉΓòÉΓòÉ
-
- This section applies only if you are using the OS/2 file system.
-
- Whenever you delete documents that have previously been indexed, you can remove
- their terms from the index:
-
- o Before deleting the documents, by placing the documents' names for removal
- in the document queue (see Putting documents in the queue for indexing)
-
- o After deleting the documents, by using cleanup.
-
- Select Cleanup to ensure that the document index contains only terms for
- existing documents. When cleanup is finished, SearchManager/2 starts document
- indexing, that is, it processes the documents in the queue.
-
- Tip If you have temporarily archived or moved documents, and do not want
- their indexed terms removed from the document index, do not do cleanup
- until you have replaced the documents again.
-
- If the Cleanup push button is unavailable:
-
- o The document index is empty
- o Indexing or merge is running
- o Cleanup is already running
- o The status of Add to queue is locked.
-
- Check the Status section of the Administration notebook for more information.
-
-
- ΓòÉΓòÉΓòÉ 8.5.5. Clearing the document index ΓòÉΓòÉΓòÉ
-
- Select Clear to remove all indexed terms from the primary and secondary parts
- of the document index.
-
- Caution The only way to undo this action is to index all the documents
- again.
-
- Like Clear for the document queue, Clear for the index is a function that you
- probably will not use very often, but it can be useful while you are trying
- out the system after installation, or if you decide to change the index type.
-
- If the Clear push button is unavailable:
-
- o The document index is empty
- o Indexing, merge, or cleanup is running.
-
- Check the Status section of the Administration notebook for more information.
-
-
- ΓòÉΓòÉΓòÉ 8.6. Document index administration ("Periodic") ΓòÉΓòÉΓòÉ
-
- Instead of manually carrying out the administration tasks Indexing, Merging,
- and Cleanup, you can use periodic administration to automate them so that they
- occur at regular intervals.
-
- This section describes:
-
- o Starting and stopping periodic administration
- o Indexing documents periodically
- o Merging the index parts periodically
- o Cleaning up periodically.
-
- To do index administration periodically, select the Periodic section of the
- administration notebook.
-
- Figure: The Periodic administration section
-
-
- ΓòÉΓòÉΓòÉ 8.6.1. Starting and stopping periodic administration ΓòÉΓòÉΓòÉ
-
- Periodic administration lets you do document indexing, index merging, and index
- cleanup, at specified intervals. You set up the three periodic administration
- tasks using the Index, Merge, and Cleanup pages of the Periodic section of the
- Administration notebook.
-
- Tip Points to keep in mind:
-
- o The three aspects of periodic administration (indexing, merge, and
- cleanup) are active only if periodic administration has been started.
-
- o The search service must be running at those times that you schedule
- periodic administration to occur. (Consider starting the search service
- automatically, and leaving it running during the periods that you have
- set for administration to occur.)
-
- To begin or end periodic administration:
-
- o Select Start or Stop.
-
-
- ΓòÉΓòÉΓòÉ 8.6.2. When periodic administration is locked ΓòÉΓòÉΓòÉ
-
- In the unlikely event that an error occurs during periodic administration, the
- Status field indicates that the administrative actions are locked. In this
- case, select Unlock to display information about why the action was stopped.
-
-
- ΓòÉΓòÉΓòÉ 8.6.3. Indexing documents periodically ΓòÉΓòÉΓòÉ
-
- At the bottom of the notebook are the three subsections for periodic
- administration. Select the Indexing section to define periodic indexing.
-
- Figure: The Periodic indexing section
-
- Tip You can define rules to prevent documents having a particular
- extension, such as .EXE, being inadvertently indexed. See Defining
- indexing rules for OS/2 document types.
-
- Select whether to process the document queue or to process selected document
- groups and the document queue.
-
- o Document queue only
-
- Specify the time interval and the maximum size for the document queue. As
- soon as the specified time or queue size has been reached, the document
- queue is processed; that is, indexing is started.
-
- Use this choice if documents are added to the document queue at irregular
- intervals rather than periodically.
-
- Tip If you want to index documents at some time in the future, but only
- once rather than periodically, specify the date and time, and then an
- interval of zero, and a queue size of zero.
-
- o Document groups
-
- If you index documents periodically, you will find indexing document groups
- more convenient than indexing individual documents. If you are continually
- changing the content of documents in a particular document group, for
- example, you can have that folder indexed periodically; you do not have to
- remember to add the changed documents to the document queue.
-
- Select Add to choose which document groups to process periodically.
-
- You can specify the starting date, time, and the interval, for each document
- group individually.
-
- Tip As with indexing the document queue, if you specify only date and
- time, and an interval of zero, periodic indexing is done only once, not
- periodically.
-
- To avoid unnecessary indexing, SearchManager/2 checks if the corresponding
- document group contains documents that have been added, or if it contains
- documents that have been updated since the last time this document group was
- indexed. Only these documents are then added to the document queue for
- indexing at the specified time.
-
- Tip This check can be made only if you have previously done periodic
- indexing of a particular document group. If you previously indexed
- "explicitly", then all the documents in this group are indexed again. So
- if you intend to index periodically, you should do it from the beginning.
-
- As soon as the specified time is reached, the names of the new or changed
- documents contained in the document groups are added to the document queue and
- the queue is processed immediately. If there are other document names in the
- queue, these documents are also processed.
-
- If SearchManager/2 finds that there are no documents to add to the queue, that
- is, no documents have been added to or changed in the document groups, any
- documents that are already in the document queue are nevertheless indexed.
-
- SearchManager/2 does not check if documents have been deleted from a document
- group. To remove obsolete indexed terms of deleted documents, refer to
- Cleaning up periodically.
-
-
- ΓòÉΓòÉΓòÉ 8.6.4. Merging the index parts periodically ΓòÉΓòÉΓòÉ
-
- To define periodic merge, select the Periodic section of the Administration
- notebook, then, at the bottom, select the Merge subsection.
-
- Figure: The Periodic merge section
-
- Ifyouchoosetouseperiodicmerge
- ,youcanpreventtheindexpartsfrommergingwhenthereareinsufficienttermsinthesecondarypart
- .( Ifyouarenotyetfamiliarwiththepartsofthedocumentindex
- ,turntoSpeedingupindexing . )
-
- The primary and secondary parts of the document index are merged only if the
- following prerequisites are both fulfilled:
-
- o The time interval set for merging is reached.
-
- Specify the starting date, time, and the interval at which you want merging
- to be done.
-
- o The size specified for the secondary part of the document index is reached.
-
- Specify the size below which you want merging prevented. The merge process
- starts only if the secondary part of the document index contains more
- documents than specified.
-
- Tip To merge the database parts only once (rather than periodically), at a
- particular time in the future, set the interval and the number of
- documents to zero.
-
-
- ΓòÉΓòÉΓòÉ 8.6.5. Cleaning up periodically ΓòÉΓòÉΓòÉ
-
- This section applies only if you are using the OS/2 file system.
-
- Whenever documents are deleted from your library or file system, you should
- remove the indexed terms for these documents from the index; it improves the
- speed of retrieval and reduces the amount of storage space taken by the index.
-
- Cleanup checks if all indexed terms refer to existing documents. If there are
- indexed terms that do not refer to a document, these indexed terms are removed
- from the document index. In addition, cleanup starts document indexing by
- processing the documents in the document queue.
-
- Tip If you have temporarily archived or moved documents, and do not want
- their indexed terms removed, stop periodic cleanup until you have replaced
- the documents again.
-
- To define periodic cleanup, select the Periodic section of the Administration
- notebook, then, at the bottom, select the Cleanup subsection.
-
- Figure: The Periodic cleanup section
-
- Specifythestartingdate ,time ,andtheintervalforperiodiccleanup
- .Ifyouspecifyonlyadateandtime ,periodiccleanupisdoneonlyonce .
-
-
- ΓòÉΓòÉΓòÉ 8.7. Displaying and changing status information ΓòÉΓòÉΓòÉ
-
- Figure: The Status section
-
- To control the status of administration tasks, select the Statusc section of
- the Administration notebook. This section lets you control the following
- activities and display their status information:
-
- o Searching
- o Adding documents to queue
- o Indexing
- o Merge
- o Cleanup.
-
- If any of the Status fields indicates that the process is locked, then an
- error has occurred. In this case, select Unlock to display information about
- the error.
-
-
- ΓòÉΓòÉΓòÉ 8.7.1. Preventing searching ΓòÉΓòÉΓòÉ
-
- Select Prevent to temporarily prevent users from searching for documents; for
- example, when you intend to clear the index.
-
- To allow searching to resume, select Permit.
-
-
- ΓòÉΓòÉΓòÉ 8.7.2. Preventing documents being added to the queue ΓòÉΓòÉΓòÉ
-
- Select Prevent to temporarily prevent users from adding document names to the
- queue; for example, when you intend to clear the queue.
-
- To allow queuing to resume, select Permit.
-
-
- ΓòÉΓòÉΓòÉ 8.7.3. Monitoring the status of indexing ΓòÉΓòÉΓòÉ
-
- Select Indexing to check the current status of indexing and when it was last
- done.
-
- The status can be Queuing, Running, Not running, or Locked.
-
-
- ΓòÉΓòÉΓòÉ 8.7.4. Monitoring the status of merge and cleanup ΓòÉΓòÉΓòÉ
-
- Select Merge or Cleanup to check their current status and when it was last
- done.
-
- The status can be Running, Not running, or Locked.
-
-
- ΓòÉΓòÉΓòÉ 8.7.5. Processing indexing messages ΓòÉΓòÉΓòÉ
-
- Indexing messages occur when, for example:
-
- o Documents cannot be indexed
- o Documents are indexed but a problem occurs
- o A dictionary cannot be found.
-
- The Indexing messages field shows the number of messages that occurred.
-
- o Select Show to display the Indexing Messages window. This window lists
- documents that could not be indexed.
-
- o Select a document from this list to display the reason why the document
- could not be indexed.
-
- o Once you have read the reason, you can select Remove selected (or Remove
- all) to remove that entry from the list.
-
-
- ΓòÉΓòÉΓòÉ 9. Searching for documents ΓòÉΓòÉΓòÉ
-
- You search for documents by specifying which terms you want to look for. You
- specify these terms in a query.
-
- In SearchManager/2, each query is an object having two main views:
-
- o A Query view: A notebook where you specify the search terms
-
- o A Results view: A folder that contains the documents found by the query.
-
- The steps in making a query are:
-
- 1. Open a search service folder
- 2. Create a query object
- 3. Open its query view
- 4. Type the search terms
- 5. Start the search.
-
- This chapter describes these steps, including how to link search terms
- together using operators, and how to enrich your search by using synonyms. It
- also describes how you can restrict a search to selected folders in the OS/2
- file system.
-
-
- ΓòÉΓòÉΓòÉ 9.1. Opening a search service folder ΓòÉΓòÉΓòÉ
-
- When you initially open a Search service folder, it contains an index object
- and one query object.
-
- To open a search service folder:
-
- o Double-click on the folder object.
-
- The open Search service folder is shown in Place the sample documents in the
- queue for indexing.
-
- The following objects are typically contained in the Search service folder:
-
- Object Description
- Index Use the pop-up menu of the Index object to:
-
- o Create a new query.
-
- o Start administration. This choice is available only for the LAN
- administrator and on a single-workstation system.
-
- o Add documents to the document queue.
-
-
- Query Query objects are those in which you write a query's search text and
- view the query's result.
-
- In a client/server environment, you may be working with several
- search services. If so, each query is associated with a particular
- search service, so it makes sense to keep the query objects,
- together with the index object, in the Search service folder. But
- you can organize your query objects into other folders, if it is
- more convenient for you.
-
- Use the pop-up menu of the Query object to:
-
- o Open a query to specify your search criteria
-
- o View the result.
-
-
- ΓòÉΓòÉΓòÉ 9.2. Creating a query object ΓòÉΓòÉΓòÉ
-
- When SearchManager/2 is first installed, a query object is provided in the
- Search service folder. To create additional Query objects, you can do one of
- the following:
-
- o Select Create another or Copy from the query's pop-up menu
-
- o Select New query from the pop-up menu of the Index object
-
- o Drag a new query object from the New query template that SearchManager/2 has
- placed in your OS/2 Templates folder.
-
- Tip Create your own query "template". To do this:
-
- 1. Create a query object.
-
- 2. Set up its defaults (see Setting defaults for a query and Sorting the
- results of a query).
-
- 3. In the General section of the query's Settings notebook, select
- Template.
-
- 4. Close the Settings notebook.
-
- You have now created a query template that you can use by dragging new
- query objects from it. Each new query has the settings that you specified.
-
- One useful setting is the sort sequence of the query's result. See Sorting
- the results of a query.
-
-
- ΓòÉΓòÉΓòÉ 9.3. Searching ΓòÉΓòÉΓòÉ
-
- (If you are working in a client/server system, you need to be logged on to the
- LAN before searching.)
-
- You make a query by entering your search text in a query notebook.
-
- To open a query notebook:
-
- o Double-click on the Query object.
-
- Figure: The Query section of the Query View notebook
-
- Each query notebook has two sections: a Query section and a Defaults section.
- The Query section is displayed when you open the notebook. To display the
- Defaults section, select the Defaults tab at the bottom left-hand side of the
- notebook.
-
-
- ΓòÉΓòÉΓòÉ 9.3.1. Specifying search terms ΓòÉΓòÉΓòÉ
-
- You can specify the following information in the Search terms fields:
-
- o A single word or a phrase, such as network
-
- o Fragments of a word or phrase using global characters, like this: net*
- Refer to Using global characters for more information.
-
-
- ΓòÉΓòÉΓòÉ 9.3.2. Selecting operators to link search terms ΓòÉΓòÉΓòÉ
-
- Select operators to specify the relation between search terms. The operators
- are or, and, and not: .
-
- or Retrieves documents that contain either of the search terms specified.
- For example:
-
- sugar
- or salt
-
- looks for documents that contain either sugar or salt, or both.
- and Retrieves documents that contain both the search terms specified. For
- example:
-
- sugar
- and salt
-
- looks for documents that contain both sugar and salt.
- not Retrieves documents that contain the first search term, but not the
- second. For example:
-
- sugar
- not salt
-
- looks for documents that contain sugar, but leaves out documents that,
- in addition, contain salt.
-
- If you use more than one operator, SearchManager/2 processes the and and the
- not operators first (in the sequence that they are given), and then it
- processes the or operators.
-
- For example:
-
- price
- and increase
- or foreign
- and airline
-
- looks for documents that contain both price and increase, and it also looks
- for documents that contain both foreign and airline. It can be thought of like
- this:
-
- (price and increase) or (foreign and airline).
-
-
- ΓòÉΓòÉΓòÉ 9.3.3. Using global characters ΓòÉΓòÉΓòÉ
-
- Global characters, otherwise known as "wild-card" characters, offer a simple
- way to make a search more flexible. They increase "recall", the number of
- documents found by a search.
-
- To set the default global characters for all queries, select Defaults in the
- pop-up menu of the SearchManager/2 folder. Any queries created subsequently
- from the Index object have the specified default global characters. Existing
- queries remain unchanged.
-
- To change the global characters for one query, use the Defaults section of its
- Query View notebook.
-
- These are the global characters provided when SearchManager/2 is first
- installed. .
-
- * An asterisk represents optional characters at the front, middle, or end of
- a search term. Here is an example of an asterisk at the front of a word:
-
- *text
-
- finds documents containing:
-
- text
- context
- hypertext
- teletext
-
- An asterisk can also represent a whole word:
-
- annual * report
-
- finds documents containing:
-
- annual budget report
- annual sales report
- annual report
-
- ? A question mark represents exactly one character in a search term:
-
- M?ll?r
-
- finds documents containing:
-
- Miller
- Myller
- Millar
-
- The question mark cannot represent a blank, nor can it stand alone to
- represent a single-character word. For example:
-
- m?ll?
-
- does not find documents containing:
-
- mill
-
-
- ΓòÉΓòÉΓòÉ 9.3.4. Searching for terms in any sequence ΓòÉΓòÉΓòÉ
-
- If you are using a precise index, SearchManager/2 searches for the terms in the
- sequence that you type them.
-
- If you are using a linguistic or a dual index however, you can select Any
- sequence; this increases recall, the number of documents likely to be found.
-
- Tip If you are using a dual index, select Inflected forms to activate the
- Any sequence choice.
-
- Multiple-word search terms: If you enter more than one word in a search term
- field, the terms are found in any sequence as long as they are in the same
- sentence. If you enter:
-
- annual report guideline
-
- the search locates documents that contain:
-
- guidelines for writing annual reports
-
- Compound search terms: If the search term belongs to a Germanic language,
- SearchManager/2 breaks a compound into its parts, and uses the parts to search
- as it would for a group of words. If you enter:
-
- Jahresbericht
-
- meaning yearly report, the search locates documents that contain:
-
- Bericht fБr das Jahr
-
- meaning report for the year.
-
-
- ΓòÉΓòÉΓòÉ 9.3.5. Searching for other (inflected) forms of terms ΓòÉΓòÉΓòÉ
-
- If you are using a precise index, SearchManager/2 searches for the terms
- exactly as you type them.
-
- If you are using a linguistic index, (the Inflected forms box contains a check
- mark), SearchManager/2 searches also for other forms of the terms, such as
- singular, plural, or a different tense.
-
- The term:
-
- drive
-
- finds documents that contain:
-
- drive
- drives
- driving
- drove
- driven.
-
- If you are using a dual index, you can choose to search for inflected forms or
- not.
-
-
- ΓòÉΓòÉΓòÉ 9.3.6. Providing synonyms for a search term ΓòÉΓòÉΓòÉ
-
- You can set up a synonym list for each search term. SearchManager/2 then
- searches for documents using not only the search term, but also the synonyms as
- alternative search terms. The synonyms are associated with a search term using
- an assumed OR operator.
-
- To associate synonyms to a particular search term:
-
- o Select Synonyms.
-
- o Select which synonyms you need for this search, then select OK.
-
- The synonyms that you selected for the specified search term are now displayed
- in the synonym list of the Query section.
-
- These synonyms are associated with the currently selected search term; you can
- select another search term and choose different synonyms for that term.
-
- Figure: The Synonyms for ... window
-
- Ifyoudonotfindaparticularsynonyminthelist ,youcanaddyourown
- .Theseadditionalsynonymsarekeptwiththequery
- ;theyarenotaddedtothepermanentlistofsynonyms .
-
- Tip If you have associated synonyms with a particular search term, but
- want to make them temporarily ineffective without deleting them, simply
- type a blank after the search term. The list of synonyms associated with
- this term is now empty.
-
- Remove the blank, and the synonyms return.
-
- If you change a search term in any way, its synonyms are cleared on the
- screen, but stored; when you reset the term, the synonyms reappear.
-
-
- ΓòÉΓòÉΓòÉ 9.3.7. Adding synonyms to the synonym list ΓòÉΓòÉΓòÉ
-
- To add synonyms to the list permanently so that they appear each time you
- display the Synonyms for window, select Permanent.
-
- The Term field contains the base form of the selected search term. If the term
- is ambiguous, that is, if there are several base forms, then these are included
- in this field.
-
- Examples:
-
- o If the search term is:
-
- lay
-
- the Term list contains:
-
- lay
- lie
-
- where lay can be the past tense of to lay or to lie.
-
- o If the search term is:
-
- building
-
- the Term list contains:
-
- building
- build
- where building can be the noun or an inflection of the verb build.
-
- To specify synonyms for a term, select it in the Term list, and add your
- synonyms to the Permanent User Synonyms list.
-
- Figure: The Permanent User Synonyms window
-
-
- ΓòÉΓòÉΓòÉ 9.3.8. An example of the use of synonyms ΓòÉΓòÉΓòÉ
-
- In the Search terms fields specified:
-
- price
- AND increase
-
- Select Synonyms for price, and choose the synonyms:
-
- charge
- tab
-
- Then select Synonyms for increase, and choose the synonyms:
-
- increment
- raise
-
- This is how SearchManager/2 now evaluates the search terms:
-
- (price OR charge OR tab )
- AND (increase OR increment OR raise)
-
-
- ΓòÉΓòÉΓòÉ 9.3.9. Selecting a language for a search term ΓòÉΓòÉΓòÉ
-
- SearchManager/2 uses a dictionary to "tune" its search to a particular
- language. Dictionaries are provided in US English and in your local language.
- See Optional IBM licensed programs for a list of other dictionaries available.
-
- You can specify a different language for each search term in the query. Select
- it from the Language drop-down list.
-
- Initially this list displays the language currently used for the query (see
- Setting defaults for a query). SearchManager/2 assumes that any synonyms
- selected for a search term are in the same language as the search term.
-
-
- ΓòÉΓòÉΓòÉ 9.3.10. Using a thesaurus for a search term ΓòÉΓòÉΓòÉ
-
- If Thesaurus End User System Toolkit/2 has been installed, there is a Thesaurus
- push button in the Query section of your Query notebook. Select Thesaurus to
- display the Thesaurus window. Here you can choose the thesaurus terms.
-
- Each term that you select from the thesaurus becomes an additional search term
- that you can link as you would any other search term using operators, like OR
- or AND. This is in contrast to synonyms, which are associated only with a
- particular search term (using an assumed OR operator).
-
- Refer to Thesaurus End User System Toolkit/2 for more information.
-
-
- ΓòÉΓòÉΓòÉ 9.4. Starting the search ΓòÉΓòÉΓòÉ
-
- When you have specified your query, select Search to start searching.
-
- SearchManager/2 displays the Results View window listing the documents found.
- For more information refer to Working with found documents.
-
- If the result contains more than 1000 documents, you are asked if you want
- SearchManager/2 to continue processing. If you answer Yes, it can nevertheless
- occur that the result displayed to you is less than 1000. This happens if you
- cannot access some documents, or if some documents have been deleted but their
- terms are still indexed.
-
-
- ΓòÉΓòÉΓòÉ 9.5. Closing a query ΓòÉΓòÉΓòÉ
-
- To close a query, double-click on the system menu symbol. You can do this
- whether or not you have run the query to make a search.
-
- Once a query has been closed, you can still view its result by selecting
- Results view from the pop-up menu of the Query object.
-
- Tip Organize your Query objects into folders, or, if you find it more
- convenient, move them directly onto the OS/2 desktop.
-
-
- ΓòÉΓòÉΓòÉ 9.6. Setting defaults for a query ΓòÉΓòÉΓòÉ
-
- Select the Defaults section of the Query View notebook to set default
- definitions for the query being processed. (To specify global defaults for all
- queries refer to Specifying standard settings (global defaults).)
-
- Figure: Query defaults
-
- Thedefaultsare :
-
- Search service
- The search service to be used; this is the one associated with the index
- that you want to search. There could be more than one to choose from if
- you are using SearchManager LAN Server/2 and Client/2 in a multiple-server
- network. To use a query for searching in several indexes, copy the query
- object and change the name specified for the search service.
-
-
- Search the entire index
- Whether to search the entire index or only the indexed terms for selected
- document groups. If you are using the OS/2 file system, you can restrict
- the search to documents contained in document groups (including subgroups)
- using the Add Document Groups window.
-
-
- Language
- The language to appear initially in the Options area of the Query section
- for each search term. For all index types, SearchManager/2 uses this
- information to do some basic linguistic processing; for linguistic indexes
- (and for dual indexes when Inflected forms is selected), SearchManager/2
- uses it to determine which dictionary to use for its advanced linguistic
- processing (described in Linguistic processing).
-
- You can override this language setting for each search term, in the
- Options area of the Query section.
-
-
- Global characters
- The global characters for the selected query. These are described in Using
- global characters.
-
-
- Display first
- How many documents can be scrolled through at one time in the Results View
- window. The maximum allowed is 300.
-
- If very many documents are found, SearchManager/2 lets you limit the
- number of document names it should process at one time. This limit ensures
- that you do not have to wait for a large result list to be displayed; the
- complete list of found documents is displayed, but in batches.
-
-
- Show progress
- Whether to display a progress indicator window during a search. The
- progress indicator lets you know how long a search is likely to take. It
- also includes a push button that lets you interrupt the search. This is
- useful when a search takes a particularly long time, and you want to
- interrupt the search so that you can repeat your query using more specific
- search terms. If you decide to use the progress indicator, keep in mind
- that the indicator itself uses system resources and can make a search take
- somewhat longer.
-
-
- ΓòÉΓòÉΓòÉ 10. Working with found documents ΓòÉΓòÉΓòÉ
-
- This section describes how to work with the results of a query, and how to use
- the Browser to display a found document, highlighting the terms that match the
- search criteria.
-
-
- ΓòÉΓòÉΓòÉ 10.1. Working with the results of a query ΓòÉΓòÉΓòÉ
-
- When you make a query, and there are documents that match the search terms,
- SearchManager/2 displays the Results View window showing the documents found.
- The status line shows the number of documents displayed.
-
- To view the result of a closed query, select Results view in the pop-up menu of
- the Query object.
-
- Figure: The Results View of a query
-
- Ifyouneedtonarrowthesearchandspecifyamoredetailedquery
- ,maketheQueryViewwindowactiveagain ,byclickingonitusingthemouse
- .Ifthiswindowiscurrentlynotvisible
- ,selectQueryViewonthesystemmenuoftheResultsViewwindow
- .TheQuerysectionofthequerynotebookisdisplayedforyoutospecifythenewormodifiedsearchcriteria .
-
- Use the menu bar of the Results View window to process the result or use the
- pop-up menu of the Document objects to process selected documents.
-
- Note: Depending on your system setup, you can edit, print, and mail the
- documents contained in the result.
-
-
- ΓòÉΓòÉΓòÉ 10.1.1. Sorting the results of a query ΓòÉΓòÉΓòÉ
-
- To sort the results of a query, select the right arrow in the Sort choice of
- the system menu. A menu of sort choices is displayed. These choices are
- provided by OS/2 for sorting the contents of a folder, but not all of them
- apply to a Results folder, whose attributes are:
-
- Title
- Location
- Description
- Last write date
- Last write time.
-
- To adjust this list for a Results View:
-
- 1. In the system menu of the query object (the object itself, or any of its
- views) select Open and then Settings.
-
- 2. In the Settings notebook, select the Sort section.
-
- 3. In the Object type list, select Found document. Now, in the Sort by
- attribute list are the attributes by which you can sort a Results folder.
-
- 4. Select all the attributes.
-
- 5. Close the Settings notebook.
-
- The next time you select Sort from the system menu of a query object, the
- correct attributes are listed.
-
- Tip To make this setting valid for all future queries, use this query as a
- "model" for others; create new queries by copying this one.
-
-
- ΓòÉΓòÉΓòÉ 10.1.2. Printing the results of a query ΓòÉΓòÉΓòÉ
-
- Select Print in the Results menu to print the list of documents resulting from
- the query. The result is printed as shown in the Details view of the Results
- View window.
-
-
- ΓòÉΓòÉΓòÉ 10.1.3. Viewing the results of a query ΓòÉΓòÉΓòÉ
-
- Use the View menu to change the view from Icon to Details view.
-
- If very many documents are found, you may notice that you cannot display all of
- their names by scrolling through the list. This is because SearchManager/2 lets
- you set a limit to the number of document names it should process at one time.
- This limit ensures that you do not have to wait for a large result list to be
- displayed. (How to specify this limit is described in Setting defaults for a
- query.)
-
- So a large number of document names are displayed to you in batches. You can
- display the next or the previous batch using the menu choices of the View menu.
-
- To see how many documents are found, display the final batch.
-
-
- ΓòÉΓòÉΓòÉ 10.2. Working with documents in the Result View ΓòÉΓòÉΓòÉ
-
- There are two ways to work with documents that have been found by a query:
-
- o Drag document objects to other objects on the OS/2 desktop. This is direct
- manipulation.
-
- o Use a document's pop-up menu.
-
- Drag/drop works only with OS/2 files.
-
-
- ΓòÉΓòÉΓòÉ 10.2.1. Direct manipulation (drag/drop) ΓòÉΓòÉΓòÉ
-
- Documents that have been found by a search can be dragged to these objects:
-
- o To an Index object. This opens the Add to queue window and lets you place
- the document in the document queue for indexing, or for removal from the
- index.
-
- o To a Printer object, for printing the document.
-
- o To an open application window that accepts dropped objects, such as OS/2's
- Enhanced Editor, or WordPerfect for OS/2.
-
- o To any Program object.
-
-
- ΓòÉΓòÉΓòÉ 10.2.2. Using a document's pop-up menu ΓòÉΓòÉΓòÉ
-
- Using normal OS/2 selection methods, you can select one or several documents
- from the result and display its pop-up menu. If the document is an OS/2 file,
- its pop-up menu contains not only the usual choices, but also the following
- additional choices provided for SearchManager/2.
-
- Open - Browser
- Select this choice to browse each document in a separate window. You may
- find this choice more convenient when you want to study and perhaps
- compare several documents at the same time. Caution is needed here,
- because each open window uses up your system's resources.
-
-
- Open - Common browser
- Select this choice to open only one window for browsing one document after
- the other. This is faster than Browser, and you may, therefore, find it
- more convenient when you want to look briefly at many documents, or save
- your system's resources.
-
-
- Add to queue
- Use this choice to add the document to the document queue for indexing or
- for removal from the index.
-
- To index several documents from the results, you have two possibilities,
- though one has a distinct disadvantage. Select the documents you want to
- have indexed, then:
-
- o Use the mouse to drag the documents to the document index object. This is
- the preferred method.
-
- o Choose Add to queue from the pop-up menu of the selected documents. This
- method is not recommended because, for each document selected, a separate
- window is opened.
-
-
- Locate original
- The document objects shown in the Results View behave like shadows of the
- original document objects. Use this choice to locate the original document
- object. This choice applies only if the document is in the OS/2 file
- system.
-
-
- Copy name
- There may be occasions when you want to use the name of a found document
- in a report, for example, or in a File - Open dialog. This choice lets you
- copy the fully qualified name of the document to the clipboard.
-
- You can use OS/2 to associate programs, such as Mail with document objects.
- You then see also these programs offered as choices in the pop-up menu of
- documents in the query result.
-
- Tip Browser choices are not available in the pop-up menu of documents in
- the result of a closed query.
-
- To browse these documents, either repeat the query (the browser choices
- are then available), or use an editor.
-
-
- ΓòÉΓòÉΓòÉ 10.3. Choosing the default Browse action ΓòÉΓòÉΓòÉ
-
- To browse a document found in the result, you can display its pop-up menu and
- select whether you want to use the Common browser or the Browser.
-
- A faster method is to double-click on the document. Initially, the default
- action for a double-click on a document in a Results view is determined by
- OS/2. However, the Options choice in the menu bar (see Working with the results
- of a query) lets you set the default to one of the two SearchManager/2 browse
- actions.
-
-
- ΓòÉΓòÉΓòÉ 10.4. Closing the Result View ΓòÉΓòÉΓòÉ
-
- To close the result, double-click on the window's system menu symbol. The
- information contained in the result is stored in the query object.
-
-
- ΓòÉΓòÉΓòÉ 10.5. Browsing found documents ΓòÉΓòÉΓòÉ
-
- Figure: Browsing a document
-
- When you browse a document, any text that matches the specified search criteria
- is highlighted. You can only view the document; you cannot change it. To change
- a document, you must use your own editor.
-
- Tip If you are using a monochrome monitor, and are having difficulty
- seeing highlighting, change the color settings through the Options choice
- in the menu bar.
-
-
- ΓòÉΓòÉΓòÉ 10.5.1. Copying text to the clipboard ΓòÉΓòÉΓòÉ
-
- To copy text to the clipboard:
-
- 1. Mark the text. Do this using either the mouse or keyboard, or using one of
- the Edit menu choices.
-
- 2. Select Copy on the Edit menu.
-
-
- ΓòÉΓòÉΓòÉ 10.5.2. Displaying an explanation of why text is highlighted ΓòÉΓòÉΓòÉ
-
- When searching for documents in a linguistic or a dual index, it might not
- always be obvious why a particular term matches the search criteria and is
- therefore highlighted. (For more information refer to Extended matching.)
-
- To display an explanation of why text is highlighted:
-
- 1. Place the cursor on the highlighted text.
-
- 2. Select Explain on the Edit menu, or select the Explain push button.
-
-
- ΓòÉΓòÉΓòÉ 10.5.3. Using extended matching ΓòÉΓòÉΓòÉ
-
- Terms that match your search criteria are highlighted in the document by the
- browser. To determine which terms to highlight, the browser makes a linguistic
- analysis of the document, in much the same way as when documents are indexed.
- This analysis occurs very quickly, and in most cases it is sufficient to find
- all occurrences of the search term.
-
- Very occasionally, there can be circumstances when the browser cannot find a
- particular term to highlight; it needs to extend its matching ability by using
- a dictionary. You can read more about this in Linguistic processing, in
- particular Extended matching.
-
- If your index is the linguistic type, this extended matching occurs
- automatically without you realizing it. First, the browser displays the
- document and highlights the matches it finds immediately; then, while you are
- reading the document, the browser makes an extensive search for more matches.
- If the browser does find further matches, a message in the information line
- informs you.
-
- To now display all matches, including those found by extended matching:
-
- Select Extended matching
-
- Extended matching is a particularly thorough process and takes time. How long
- it takes depends on factors like the size of the document, the language in
- which the document is written, and the speed of your workstation; it can range
- from a few seconds to many minutes.
-
-
- ΓòÉΓòÉΓòÉ 10.5.4. Processing hypermedia references ΓòÉΓòÉΓòÉ
-
- Documents can contain hypermedia references. These are points in a document
- from which other information, such as graphics, images, audio, or video, can be
- called.
-
- For example, a document could include a hypermedia reference to a piece of
- music. By selecting the reference, you could play the music using a CD player
- attached to the workstation.
-
- SearchManager/2 includes an application programming interface (API) that makes
- it possible for you to write the programs that process whatever media your
- documents contain. This API is described in IBM SearchManager/2: Programming
- Interfaces.
-
- An example of how you could implement hypermedia references is included in the
- sample documents. One of these documents contains a reference to a picture of a
- hot-air balloon:
-
- 1. Search in the sample documents for the term hot air.
-
- 2. Browse the found document.
-
- 3. Place the cursor on the hypermedia reference.
-
- 4. Select Hypermedia on the Edit menu, or select the Hypermedia push button.
-
- To set the color in which hypermedia references are displayed, use the Options
- choice in the menu bar.
-
-
- ΓòÉΓòÉΓòÉ 10.5.5. Locating words while browsing a document ΓòÉΓòÉΓòÉ
-
- While you are browsing a document, you can move easily from match to match,
- find a particular word, or even find the next occurrence of a term that you
- have selected in the text.
-
- To do this, use the Next match and Previous match push buttons, or the choices
- in the Locate menu in the menu bar.
-
-
- ΓòÉΓòÉΓòÉ 10.5.6. Setting fonts and colors ΓòÉΓòÉΓòÉ
-
- To set fonts and colors for the Browser window, use the Options menu. The fonts
- are monospaced.
-
-
- ΓòÉΓòÉΓòÉ 11. Installation ΓòÉΓòÉΓòÉ
-
- SearchManager/2 is designed for an OS/2 local area network (LAN) client/server
- environment, and for OS/2 single-workstation systems. In a typical LAN
- environment, a SearchManager/2 installation comprises a server workstation (or
- possibly several) on which SearchManager/2's "search service" program is
- installed, and several client workstations with access to the search service.
-
- The IBM SearchManager LAN Server/2 and Client/2 product includes not only the
- software for installation at the server workstation, but also the software for
- client workstations.
-
- The SearchManager/2 single-workstation product contains the software for
- installation on a single workstation; the search service program and the
- "client" program are installed together in one workstation.
-
- This chapter tells you:
-
- o How to plan for the installation
-
- o What hardware and software you need
-
- o How to install SearchManager LAN Server/2 and Client/2
-
- o How to install the SearchManager/2 single-workstation product
-
- o How to automate the installation if you need to carry it out at several
- workstations.
-
- Before you start installation: Read the READ.ME file provided on the
- Installation diskette to find out the latest information available after this
- book was written.
-
-
- ΓòÉΓòÉΓòÉ 11.1. Installation planning ΓòÉΓòÉΓòÉ
-
- This section helps you to plan:
-
- o For good performance
- o Access to documents
- o For working with Lotus Notes
- o Which dictionaries to install
- o Using SearchManager LAN Server/2 and Client/2 in a Novell** NetWare**
- network.
-
-
- ΓòÉΓòÉΓòÉ 11.1.1. Planning for performance ΓòÉΓòÉΓòÉ
-
- In a system where large volumes of data are being indexed or searched, good
- system performance is particularly important. Here are some general points to
- help you ensure that SearchManager/2 operates quickly and efficiently:
-
- o Processor: You can install SearchManager/2 on workstations having an Intel**
- 386 processor, but you may not be satisfied with its processing speed in a
- client/server system. A server workstation, in particular, should have a
- powerful processor; recommended is at least a 486 processor, running at 25
- Mhz.
-
- o Disk speed: If possible, the hard disk should have an average access time of
- at least 12 ms.
-
- o Memory and size of hard disk: This depends on the number, size and type of
- documents you have in your library system. Read the sections Hardware for
- SearchManager LAN Server/2 and Client/2 and Hardware for SearchManager/2, in
- particular the information concerning memory, permanent disk space, and
- temporary disk space.
-
- o Other applications: For single-workstation systems, and for server
- workstations in a client/server system, be cautious about running additional
- applications simultaneously. This increases the system load and disk
- input/output, which in turn slows down all applications running on that
- workstation. If possible, distribute applications to other server
- workstations.
-
- o Environment: When you install SearchManager/2 you can choose where to
- install its various components. In a client/server system, prefer to install
- executable files locally. If you install them on remote drives, it increases
- the traffic that has to be handled by the local area network, and can
- therefore have an adverse effect on the time taken for loading DLLs and
- other required data, like temporary work files.
-
-
- ΓòÉΓòÉΓòÉ 11.1.2. Planning access to documents ΓòÉΓòÉΓòÉ
-
- For a single-workstation installation, there is no restriction on where
- documents are located.
-
- For a client/server installation, the only restriction on the location of
- documents is that they must be accessible by the SearchManager/2 client for
- retrieval, and by the SearchManager/2 server for indexing.
-
- If SearchManager LAN Server/2 is installed on a file server, and the documents
- on this file server are to be indexed at the file server, then you must connect
- to the documents using the NET USE command. In this way, you ensure that
- Universal Naming Convention (UNC) names are used, so that the clients can
- retrieve the documents even if they are connected to them using different drive
- letters. (If documents on the file server are indexed from a remote client, UNC
- names are used automatically.)
-
-
- ΓòÉΓòÉΓòÉ 11.1.3. Planning to work with Lotus Notes ΓòÉΓòÉΓòÉ
-
- These are points to take into consideration if you intend to work with Lotus
- Notes documents:
-
- o OS/2 file system: If you plan to work with Lotus Notes documents and OS/2
- file system documents, each system requires its own installation of
- SearchManager LAN Server/2 on a separate server workstation.
-
- o Client: A Lotus Notes Version 3 client must be installed.
-
- The access authorization of this client must have the highest priority
- needed to satisfy all database requests, such as reading all relevant
- databases, views, forms, and documents.
-
- The client must be in the Access Control Lists of the databases to be
- indexed.
-
- o Password: Lotus Notes initialization routines ask for a password to be
- entered, if required by the Lotus Notes user ID. This can cause a
- password-prompt window to open at an unattended server workstation.
-
- To avoid this:
-
- - Provide a Lotus Notes user ID for the SearchManager/2 server workstation,
- without a password (password cleared). Ensure that this user ID is
- assigned an access level of at least Editor.
-
- - To maintain security, install (or enable) a keyboard-protection program,
- such as OS/2's Lockup.
-
-
- ΓòÉΓòÉΓòÉ 11.1.4. Deciding which dictionaries to install ΓòÉΓòÉΓòÉ
-
- Dictionaries are used to support the linguistic processing of documents in
- different languages during indexing and retrieval.
-
- SearchManager/2 provides two dictionaries, one for US English and one for your
- local language.
-
- Tip You should install at least one dictionary for the main language of
- the documents to be searched. If you do this:
-
- o You are able to use all the linguistic functions described in Linguistic
- processing.
-
- o Irrelevant words ("stop words") are excluded from the index.
-
- o You have a synonym list from which you can choose additional search
- terms.
-
- If you initially decide not to install dictionaries, but then install
- dictionaries later, for consistent search results you will need to index all
- documents again. This is true regardless of the index type; even in a precise
- index, dictionaries are used for processing stop words.
-
- If you need dictionaries in addition to those that are included in the
- SearchManager/2 package, refer to Optional IBM licensed programs.
-
-
- ΓòÉΓòÉΓòÉ 11.1.5. Deciding where to install dictionaries ΓòÉΓòÉΓòÉ
-
- In a single-workstation system, this question does not arise, because the
- dictionaries can be stored only on that workstation.
-
- In a client/server installation, if the server workstation on which
- SearchManager LAN Server/2 is installed is a file server, you could save space
- by installing the dictionaries only at the server workstation; the client can
- make use of the server's dictionaries by defining a path to them. You could
- also install dictionaries on a file server that is neither a SearchManager/2
- client or server workstation.
-
- However, you are recommended to install dictionaries at the server workstation,
- and at each client workstation. Installing dictionaries locally can result in a
- substantial improvement in performance.
-
- Dictionaries installed at the server workstation are used by the search service
- while indexing and searching. They are used for:
-
- o Reduction to base form
- o Decomposition
- o Stop-word filtering
- o Part-of-speech filtering
- o Dehyphenation.
-
- Dictionaries installed at the client workstation are used for:
-
- o Finding synonyms when making a query
- o Extended matching to find obscure matches.
-
-
- ΓòÉΓòÉΓòÉ 11.1.6. Defining a search service when you have no network ΓòÉΓòÉΓòÉ
-
- SearchManager LAN Server/2 and Client/2 normally requires a network of course,
- but, for test purposes, you could install the client and the server components
- on a single workstation without networking software. In this case, these are
- the parameters you must specify when defining the search service:
-
- o Service name: SM2
-
- o Machine name: SM2
-
- o Communication type: OS/2 named pipes.
-
-
- ΓòÉΓòÉΓòÉ 11.1.7. Using SearchManager/2 in a Novell NetWare network ΓòÉΓòÉΓòÉ
-
- SearchManager LAN Server/2 and Client/2 can be used in a Novell NetWare 3.11
- network using named pipes. All workstations in the network must be configured
- as NetWare requesters for OS/2. In this case the SearchManager LAN Server/2
- workstation need not be configured as an OS/2 LAN Server (using IBM OS/2 LAN
- Server 2.0 or 3.0), but it must be configured as a NetWare named pipes server.
-
- These are the points to consider if you are planning to use SearchManager LAN
- Server/2 and Client/2 in a Novell NetWare network without IBM OS/2 LAN Server.
-
- No remote administration: Administration can be carried out only at the
- workstation where SearchManager LAN Server/2 has been installed. Remote
- administration can be carrried out only by an administrator of IBM OS/2 LAN
- Server.
-
- If documents are on a NetWare file server: If the documents to be indexed are
- on a NetWare file server and you are using the OS/2 library services (OS/2 file
- system) provided with SearchManager/2:
-
- o To enable SearchManager/2 clients to find the documents that have been
- indexed, SearchManager LAN Server/2 and its clients must use identical drive
- letters when drives are mapped to directories on a NetWare file server.
-
- o Documents cannot be added to the queue from a remote SearchManager/2 client
- workstation, because drives that are mapped to directories on a NetWare file
- server are treated like local drives.
-
- o For the same reason, only the whole document index can be searched from a
- remote SearchManager/2 client workstation, that is, a search cannot be
- restricted to document groups.
-
- o On the workstation where SearchManager LAN Server/2 has been installed, the
- document group windows (Add to queue, restrict search to document groups)
- show not only mapped drives and their subdirectories, but also the documents
- that are located on a mapped drive and in its subdirectories.
-
-
- ΓòÉΓòÉΓòÉ 11.1.8. Using SearchManager/2 in a TCP/IP network ΓòÉΓòÉΓòÉ
-
- You can use SearchManager/2 in a TCP/IP network in conjuction with IBM OS/2 LAN
- server 2.0 or 3.0. However, these are the points to consider if you are
- planning to use TCP/IP without IBM OS/2 LAN Server.
-
- No remote administration: Administration can be carried out only at the
- workstation where SearchManager LAN Server/2 has been installed. Remote
- administration can be carrried out only by an administrator of IBM OS/2 LAN
- Server.
-
- If documents are on a Network File System server: If the documents to be
- indexed are on a Network File System (NFS) server, and you are using the OS/2
- library services (OS/2 file system) provided with SearchManager/2:
-
- o To enable SearchManager/2 clients to find the documents that have been
- indexed, SearchManager LAN Server/2 and its clients must mount directories
- on an NFS server using identical drive letters.
-
- o Documents cannot be added to the queue from a remote client workstation,
- because mounted directories on an NFS server are treated like local drives.
-
- o For the same reason, only the whole document index can be searched from a
- remote SearchManager/2 client workstation, that is, a search cannot be
- restricted to document groups.
-
-
- ΓòÉΓòÉΓòÉ 11.2. Installing IBM SearchManager LAN Server/2 and Client/2 ΓòÉΓòÉΓòÉ
-
- This section describes how to install SearchManager LAN Server/2 and Client/2.
- If you are installing the SearchManager/2 single-workstation system, refer to
- Installing IBM SearchManager/2 single workstation.
-
- If the SearchManager/2 single-workstation system is already installed, then you
- must delete it before proceeding; SearchManager LAN Server/2 and
- SearchManager/2 single-workstation cannot be installed together on the same
- workstation. Remember to save your index files before doing so; Deleting
- components of SearchManager/2 describes how to do this.
-
- Tip Before proceeding, check that you have the diskettes for SearchManager
- LAN Server/2 and Client/2, and not those for the SearchManager/2
- single-workstation system.
-
- Follow these steps to install the server AND the client components on the
- server workstation, and then repeat them to install the client component at
- each client workstation.
-
- If you need to install components of SearchManager LAN Server/2 and Client/2
- on several workstations, you may want to do the installation from the LAN, or
- even automate the installation. If so, refer to Installing SearchManager/2
- without diskettes.
-
-
- ΓòÉΓòÉΓòÉ 11.2.1. Installation configurations for LAN systems ΓòÉΓòÉΓòÉ
-
- The following figure illustrates a client/server system having one server.
-
- Figure: SearchManager/2 in a single-server network
-
- The search service, installed on the server workstation, handles the indexing
- and retrieval tasks. The document index must be located on the same server
- workstation on which the search service is installed. Each client has a search
- service object through which it accesses the search service installed at the
- server.
-
- The documents themselves could be installed at the server workstation, or at a
- file server to which the server workstation and the clients have access.
-
- If the documents are located at several file servers, the server workstation at
- which the search service is installed needs access to all the file servers. The
- clients need access to those file servers that contain the documents they are
- authorized to work with.
-
- The following figure illustrates a client/server system having more than one
- server. A search service is installed at each server workstation. Each search
- service handles the indexing and retrieval tasks for one document index. Each
- client has one search service object for each document index that it needs to
- work with.
-
- Figure: SearchManager/2 in a two-server network
-
-
- ΓòÉΓòÉΓòÉ 11.2.2. Hardware for SearchManager LAN Server/2 and Client/2 ΓòÉΓòÉΓòÉ
-
- This is the hardware that you need to install SearchManager LAN Server/2 and
- Client/2.
-
- For a server workstation:
-
- o IBM Personal System/2* or equivalent with an Intel 386 processor (or
- equivalent) or higher
-
- o VGA, 8514, XGA or SVGA (IBM) display adapter
-
- o 4MB main storage, in addition to the memory needed for the operating system
- and any other active applications
-
- o Diskette drive: 1.44MB, 3.5"
-
- o Disk space:
-
- Server program 6.5MB
- and client program
- Each dictionary between 0.5MB and 3.0MB
-
- Toolkit and 0.5MB
- programming samples
- Online manual 0.5MB
-
- Document index See below.
-
- o Mouse (recommended).
-
- Disk space for indexing:
-
- You need permanent disk space for the document index, plus temporary disk
- space for merging the index parts and temporary disk space for indexing. The
- temporary disk space requirement is not additive, because merging and indexing
- do not run at the same time.
-
- Unfortunately, the disk space calculation is not straight forward, so take
- time to understand it, and avoid running into disk space problems later.
-
- o Permanent disk space for the document index
-
- This depends on many factors, such as the number of graphics that your
- documents contain; the more graphics there are, the less space you need for
- the index.
-
- It depends on the type of documents you index; documents from Ami Pro and
- similar word-processing products need relatively less space because much of
- their content consists of control characters rather than text.
-
- It also depends on the number of unique words or numbers that your documents
- contain; documents containing lists of sequential numbers take up relatively
- more space because each number is unique and is indexed separately.
-
- Lastly, it depends, of course, on the number and size of your documents.
-
- As a guideline, reserve disk space for about 0.7 times the size of the
- documents to be indexed. For example, you would need about 7MB of disk space
- for a document index for 10MB of documents. Then experiment by indexing a
- sample batch of typical documents.
-
- o Temporary disk space for merging
-
- For merging, reserve about 2.5 times the size of the primary part and the
- secondary part of the document index.
-
- o Temporary disk space for indexing
-
- As with permanent disk space, the amount of temporary disk space that you
- need for indexing can vary widely. It depends on the size of your secondary
- index, and on the size of the documents to be indexed.
-
- Calculate it as follows:
-
- 1. Take the current size of the secondary part of the document index.
-
- 2. Add about 0.7 times the size of the documents to be indexed.
-
- 3. Add to this again about 0.7 times the documents to be indexed. If this
- last calculation comes to more that 10MB, then use the figure 10MB.
-
- For a client workstation:
-
- o IBM Personal System/2 or equivalent with an Intel 386 processor (or
- equivalent) or higher
-
- o VGA, 8514, XGA or SVGA (IBM) display adapter
-
- o 2MB main storage, in addition to the memory needed for the operating system
- and any other active applications
-
- o Diskette drive: 1.44MB, 3.5"
-
- o Disk space:
-
- Client program 5MB
- Each dictionary between 0.5MB and 3.0MB
-
- Toolkit and 0.5MB
- programming samples
- Online manual 0.5MB
-
- o Mouse (recommended).
-
-
- ΓòÉΓòÉΓòÉ 11.2.3. Software for SearchManager LAN Server/2 and Client/2 ΓòÉΓòÉΓòÉ
-
- This is the software that you need to install SearchManager LAN Server/2 and
- Client/2.
-
- For a server workstation:
-
- o IBM OS/2 Version 2 (5621-077), CSD level XR06055, or above.
- o IBM OS/2 LAN Server Version 2.0, CSD level IP06030 or Version 3.0
- (5621-253).
-
- You can use OS/2 LAN Server alone, or together with one of these:
-
- - IBM TCP/IP Version 1.2.1 for OS/2 (5798-RXW)
-
- - Novell NetWare Version 3.11.
-
- If you use OS/2 LAN Server alone, the workstation on which SearchManager LAN
- Server/2 is installed must be configured as a LAN server.
-
- It is also possible to use TCP/IP or Novell NetWare without OS/2 LAN Server.
- See Using SearchManager/2 in a TCP/IP network and Using SearchManager/2 in a
- Novell NetWare network for details.
-
- For a client workstation:
-
- o IBM OS/2 Version 2 (5621-077), CSD level XR06055 or above.
- o IBM OS/2 Requester from IBM OS/2 LAN Server Version 2.0 CSD level XR06030,
- or Version 3.0 (5621-253).
- o IBM TCP/IP Version 1.2.1 for OS/2 (5798-RXW), optional.
-
- The workstation on which Client/2 is installed must be configured as a LAN
- requestor if you are using OS/2 LAN Server at the server workstation.
-
-
- ΓòÉΓòÉΓòÉ 11.2.4. Starting the client and server installation ΓòÉΓòÉΓòÉ
-
- Essential steps Installation consists not only in running the installation
- program to install the software; there are also some configuration steps.
-
- Use this checklist to ensure that you carry out all the installation
- steps.
-
- 1. Read the section that describes installation planning, in particular
- Planning access to documents.
-
- 2. At the server workstation, install both the server and the client
- components.
-
- Be sure that your documents are located on a LAN file server that is
- accessible by all clients.
-
- During installation, install the sample documents to a path that
- client workstations can access.
-
- 3. Set up the SearchManager/2 global defaults.
-
- 4. Set up the search service.
-
- 5. At the client workstations, install the client component.
-
- 6. At each client workstation, create a search service object through
- which the client can access the search service installed at the server
- workstation.
-
- 7. When you have finished the installation at a particular workstation,
- verify the installation.
-
- 8. Index your documents. If you plan to use periodic indexing, do your
- initial indexing using periodic indexing.
-
- If you have already installed SearchManager LAN Server/2 or Client/2, the
- installation program recognizes this and presents suitable options for update
- or reinstallation. A delete option is also available.
-
- The following components can be installed:
-
- IBM SearchManager Client/2
- The client component for working with the IBM SearchManager LAN
- Server/2 server component. You must install the client component
- also on the server workstation.
-
-
- SearchManager LAN Server/2
- The server component and sample documents for installation on a
- server workstation. Install these on a LAN server or on a TCP/IP
- workstation. The sample documents are used in verifying the
- installation.
-
-
- Online Documentation
- This book, IBM SearchManager/2: Retrieval and Indexing, is included
- not only as a manual, but also as online information. You can view
- it by double-clicking on the Retrieval and Indexing object in the
- SearchManager/2 folder.
-
-
- Dictionaries
- Monolingual dictionaries and stop-word lists for two languages.
-
-
- Toolkit and programming samples
- Files for the SearchManager/2 Library Services sample, and for the
- SearchManager/2 Retrieval API sample. These are described in IBM
- SearchManager/2: Programming Interfaces.
-
- Figure: Components of SearchManager LAN Server/2 and Client/2
-
- To install SearchManager LAN Server/2 and Client/2:
-
- 1. Start an OS/2 session.
-
- 2. Insert the Installation diskette into a diskette drive.
-
- 3. Make sure that the current drive is the drive containing the installation
- program; for example, A: if you install the program from drive A.
-
- 4. Type SM2INST at the OS/2 command prompt and press Enter.
-
- 5. Follow the instructions on the screens; select Help for more details.
-
- 6. When you have finished, restart your operating system.
-
- 7. Double-click on the Initialize SearchManager/2 object to start the process
- that completes the installation.
-
-
- ΓòÉΓòÉΓòÉ 11.2.5. How SearchManager LAN Server/2 and Client/2 modifies your CONFIG.SYS ΓòÉΓòÉΓòÉ
-
- These are the changes that SearchManager/2 makes to your CONFIG.SYS file:
-
- o The directory in which the dynamic link libraries (DLLs) are installed is
- added to:
-
- LIBPATH
-
- o The directory in which the product is installed is added to:
-
- SET PATH
- SET DPATH
- SET HELP
-
- o The following environment variables are added:
-
- EHSDATA: directory for the SearchManager/2 data files (on a server
- workstation also for the index files)
- EHSWORK: directory for temporary files
- EHSNLPS: directory for dictionaries
- TMP (if not already present): for temporary files, pointing to the same
- directory as EHSWORK.
-
-
- ΓòÉΓòÉΓòÉ 11.2.6. Specifying standard settings (global defaults) ΓòÉΓòÉΓòÉ
-
- Once you have installed SearchManager/2, set up the Global defaults. These are
- the settings that you want to apply to all queries. Changing them affects only
- new queries and existing queries that you open after making the change.
-
- You can set:
-
- o The language used on windows.
-
- o The language used for queries.
-
- You can change the language selection for individual search terms and
- queries. Refer to Selecting a language for a search term, and Setting
- defaults for a query.
-
- o Global ("wild-card") characters used when specifying queries.
-
- You can also select global characters for individual queries. Refer to Setting
- defaults for a query.
-
- To set the global defaults:
-
- o Select Defaults on the pop-up menu of the SearchManager/2 folder.
-
-
- ΓòÉΓòÉΓòÉ 11.2.7. Defining the local search service at the server ΓòÉΓòÉΓòÉ
-
- The SearchManager/2 search service provides you with indexing and retrieval
- services, such as starting a search, and adding documents to the queue for
- indexing. At the workstation where the search service is installed, it also
- handles the indexing of documents.
-
- On a client/server system, the search service is installed at the server
- workstation, but each client can use the service through the search service
- object in the SearchManager/2 folder.
-
- Figure: A search service in a client/server system
-
- If a client is connected to more than one server workstation, it can have one
- search service object for each server on which SearchManager LAN Server/2 is
- installed. Installation configurations for LAN systems shows a possible
- configuration.
-
- In a client/server system, you set up the search service to ensure that it runs
- most efficiently for your system, and that the client and the server work
- together using the same parameters.
-
- To set up the server's search service:
-
- 1. Open the SearchManager/2 folder.
-
- 2. Select Settings on the pop-up menu of the Search service folder.
-
- 3. Select the Definitions section.
-
- Figure: Defining a search service
-
- Tip Changes to the search service settings are active only the next time
- you start it. To activate new settings immediately, stop the search
- service, then restart it.
-
- In the Definitions section of the Search service folder you set up:
-
- o The name of the service
- o The name of the server workstation and the communication type
- o Some service parameters.
-
- Search service name: This is the name that you want to give to the local
- search service. It is needed also by the client to access the search service.
- It does not have to be the same name that you give to the search service
- object on the OS/2 desktop.
-
- Define search service:
-
- Machine name
- This is the identification of the server workstation where the
- search service is installed. It is dependent on the communication
- type:
-
- Named pipes: Enter the LAN address of this workstation. You can
- display this by entering
-
- net name
-
- in an OS/2 session.
-
- TCP/IP: Enter the TCP/IP host name of this workstation if the name
- is not longer than 15 characters. Otherwise, enter the TCP/IP dotted
- decimal address of this workstation.
-
-
- Communication type
- OS/2 named pipes, or TCP/IP.
-
-
- Port
- Required for TCP/IP only.
-
- Service parameters: These are tuning parameters for making your system work
- more efficiently.
-
- Max. parallel service tasks
- The maximum number of tasks to be handled at the same time by the
- search service. Service tasks are, for example, starting a search,
- adding documents to the queue, or administration tasks.
-
- The higher you set this value, the slower your system becomes. On
- the other hand, if you set this value too low, then, at busy times,
- some clients may not be able to search.
-
- The value set initially is 6.
-
- Tip If the server has, say, 30 clients that are not likely to be searching
- at the same time, try setting this value to about 10.
-
-
- Service tasks kept available
- The number of search service tasks to be kept available. This is the
- number of tasks started in advance to handle future requests.
-
- For good performance, a client making a request should always find a
- service task ready to handle its request. However, having too many
- service tasks open unnecessarily means making inefficient use of the
- service.
-
- The specified number must be lower or equal to the number of search
- service tasks accepted.
-
- The value set initially is 3.
-
-
- Time limit seconds
- The time that a search service task is allowed to run. After this
- time, the service task assigned to handle a specific request, like a
- search, is terminated.
-
- Set this value to the longest reasonable search time. It prevents
- unreasonably long searches from reducing the overall system
- performance.
-
- It also terminates inactive tasks, such as when a query notebook is
- left open after a search.
-
- The recommended minimum value is 1200 seconds, which is the value
- set initially.
-
-
- Task cleanup every ms
- How frequently you want SearchManager/2 to check whether new tasks
- have to be started to handle future requests, or to check whether to
- remove surplus tasks to ease the system load.
-
- It is best not to change the initial value which is 500ms.
-
-
- ΓòÉΓòÉΓòÉ 11.2.8. Defining a search service in a Novell NetWare network ΓòÉΓòÉΓòÉ
-
- These are the points to note when setting up the search service for a server in
- a Novell NetWare network.
-
- o Machine name: Enter the name of the NetWare named pipes server workstation.
- If you are using Novell NetWare in addition to IBM OS/2 LAN Server 2.0 or
- 3.0, then this field initially shows the machine name of the IBM OS/2 LAN
- Server.
-
- o Communication method: Set the communication method to OS/2 Named Pipes
-
-
- ΓòÉΓòÉΓòÉ 11.2.9. Accessing a remote search service from a client ΓòÉΓòÉΓòÉ
-
- When you install Client/2 on a client workstation, a New remote search service
- object is created in the in the SearchManager/2 folder.
-
- You must set up the client's Search service object to contain the server's name
- and address.
-
- To set up a search service object for a client:
-
- 1. Open the Settings view of the New remote search service object.
-
- 2. Select the Definitions section.
-
- In the Definitions section of the Search service folder you set up:
-
- o The name of the remote search service
- o The address of the search service.
-
- These parameters must be identical to those specified at the server
- workstation.
-
- Service name: The name of the remote search service you want to work with.
-
- Tip Be sure that this name is identical in every way to the name specified
- at the server workstation.
-
- (It does not have to be the same name that you give to the search service
- object on the OS/2 desktop.)
-
- Service address:
-
- Machine name
- The identification of the server workstation where the search
- service is installed.
-
-
- Communication type
- The communication type of the server workstation where the search
- service is installed (OS/2 named pipes, or TCP/IP).
-
-
- Port
- The same as on the server workstation (required for TCP/IP only).
-
-
- ΓòÉΓòÉΓòÉ 11.2.10. Creating additional search service objects for a client ΓòÉΓòÉΓòÉ
-
- If additional search services are installed on other server workstations, then,
- at the client workstation, you can create more search service objects to
- provide access to each search service.
-
- To create an additional search service object for a client:
-
- 1. Click on the SearchManager/2 folder using the right-hand mouse button to
- display its pop-up menu.
-
- 2. Select New remote service. A new remote service object appears in the
- SearchManager/2 folder.
-
- 3. Open the Settings view of the new remote search service object.
-
- 4. Select the Definitions section.
-
- 5. Follow the steps described in Accessing a remote search service from a
- client.
-
-
- ΓòÉΓòÉΓòÉ 11.2.11. Verifying the SearchManager LAN Server/2 and Client/2 installation ΓòÉΓòÉΓòÉ
-
- When you have finished installing, you should verify the installation by
- indexing the sample documents and then carrying out a simple search.
-
- For a server workstation: Installation verification consists of the following
- steps:
-
- 1. Place sample documents in the document queue for indexing
- 2. Index the sample documents
- 3. Search for terms contained in the documents
- 4. Browse documents to see the terms found.
-
- Follow the description of these steps in Getting started, beginning at A brief
- look at the indexing process.
-
- For a client workstation: Installation verification consists of the following
- steps:
-
- o Search for terms contained in the documents
- o Browse documents to see terms found.
-
- Follow the description of these steps in Getting started, beginning at Making
- your first search.
-
-
- ΓòÉΓòÉΓòÉ 11.3. Installing IBM SearchManager/2 single workstation ΓòÉΓòÉΓòÉ
-
- This section describes installing SearchManager/2 on a single workstation. If
- you are installing SearchManager LAN Server/2 and Client/2 (client/server
- system), turn to Installing IBM SearchManager LAN Server/2 and Client/2.
-
- If SearchManager LAN Server/2 is already installed, then you must delete it
- before proceeding; SearchManager LAN Server/2 and the SearchManager/2
- single-workstation system cannot be installed on the same workstation. Remember
- to save your index files before doing so, Deleting components of
- SearchManager/2 describes how to do this.
-
- Tip Before proceeding, check that you have the diskettes for
- SearchManager/2 single-workstation, and not those for SearchManager LAN
- Server/2 and Client/2.
-
-
- ΓòÉΓòÉΓòÉ 11.3.1. Hardware for SearchManager/2 ΓòÉΓòÉΓòÉ
-
- o IBM Personal System/2 or equivalent with Intel processor 80386 or
- equivalent/higher
-
- o VGA, 8514, XGA or SVGA (IBM) display adapter
-
- o 4MB main storage, in addition to the memory needed for the operating system
- and any other active applications
-
- o Diskette drive: 1.44MB, 3.5"
-
- o Disk space:
-
- SearchManager/2 program 6.0MB
- Each dictionary between 0.5MB and 3MB
-
- Toolkit and 0.5MB
- programming samples
- Online manual 0.5MB
-
- Document index See Hardware for SearchManager LAN
- Server/2 and Client/2
-
-
- ΓòÉΓòÉΓòÉ 11.3.2. Software for SearchManager/2 ΓòÉΓòÉΓòÉ
-
- o IBM OS/2 Version 2 (5621-077), CSD level XR06055 or above.
-
-
- ΓòÉΓòÉΓòÉ 11.3.3. Starting the SearchManager/2 installation ΓòÉΓòÉΓòÉ
-
- An installation program guides you and prompts you for the components to be
- installed and other relevant details. If you have already installed
- SearchManager/2, the installation program recognizes this and presents suitable
- options for update or reinstallation. A delete option is also available.
-
- Essential steps Installation consists not only in running the installation
- program to install the software; there are also some configuration steps.
-
- Use this checklist to ensure that you carry out all the installation
- steps.
-
- 1. Read the section that describes installation planning.
-
- 2. Install the SearchManager/2 single-workstation system.
-
- 3. Set up the SearchManager/2 global defaults.
-
- 4. Set up the search service.
-
- 5. Verify the installation.
-
- 6. Index your documents. If you plan to use periodic indexing, do your
- initial indexing using periodic indexing.
-
- The following components can be installed:
-
- SearchManager/2
- The SearchManager/2 single-workstation component and sample
- documents used in verifying the installation.
-
-
- Online Documentation
- This book, IBM SearchManager/2: Retrieval and Indexing, is included
- not only as a manual, but also as online information so that you can
- view it by double-clicking on the Retrieval and Indexing object in
- the SearchManager/2 folder.
-
-
- Dictionaries
- Monolingual dictionaries and stop-word lists for two languages.
-
-
- Toolkit and programming samples
- Files for the SearchManager/2 Library Services sample, and for the
- SearchManager/2 Retrieval API sample. These are described in IBM
- SearchManager/2: Programming Interfaces.
-
- To install the SearchManager/2 single-workstation system:
-
- 1. Start an OS/2 session.
-
- 2. Insert the Installation diskette into a diskette drive.
-
- 3. Make sure that the current drive is the drive containing the installation
- program. For example, A: if you install the program from drive A.
-
- 4. Type SM2INST at the OS/2 command prompt and press Enter.
-
- 5. Follow the instructions on the windows; select Help for more details. The
- main installation panel is similar to the one shown in Starting the client
- and server installation.
-
- 6. When you have finished, restart your operating system.
-
- 7. Double-click on the Initialize SearchManager/2 object to start the process
- that completes the installation.
-
-
- ΓòÉΓòÉΓòÉ 11.3.4. How SearchManager/2 modifies your CONFIG.SYS ΓòÉΓòÉΓòÉ
-
- These are the changes that SearchManager/2 makes to your CONFIG.SYS file:
-
- o The directory in which the dynamic link libraries (DLLs) are installed is
- added to:
-
- LIBPATH
-
- o The directory in which the product is installed is added to:
-
- SET PATH
- SET DPATH
- SET HELP
-
- o The following environment variables are added:
-
- EHSDATA: directory for the index and other data files
- EHSWORK: directory for temporary files
- EHSNLPS: directory for dictionaries
- TMP (if not already present): for temporary files, pointing to the same
- directory as EHSWORK.
-
-
- ΓòÉΓòÉΓòÉ 11.3.5. Specifying standard settings (global defaults) ΓòÉΓòÉΓòÉ
-
- Once you have installed SearchManager/2, set up the Global defaults. These are
- the settings that you want to apply to all queries.
-
- Changing them affects only new queries and existing queries that you open after
- making the change.
-
- You can set:
-
- o The language used for queries.
-
- You can change the language selection for individual search terms and
- queries. Refer to Selecting a language for a search term, and Setting
- defaults for a query.
-
- o Global ("wild-card") characters used when specifying queries.
-
- You can also select global characters for individual queries. Refer to Setting
- defaults for a query.
-
- To set the global defaults:
-
- o Select Defaults on the pop-up menu of the SearchManager/2 folder.
-
-
- ΓòÉΓòÉΓòÉ 11.3.6. Setting up the search service for a single-workstation ΓòÉΓòÉΓòÉ
-
- The SearchManager/2 search service provides you with indexing and retrieval
- services, such as starting a search, and adding documents to the queue for
- indexing. It also handles the indexing of documents. You set up a search
- service so that it runs most efficiently for your system.
-
- Figure: The search service on a single-workstation system
-
- To set up the search service:
-
- 1. Open the SearchManager/2 folder.
-
- 2. Select Settings on the pop-up menu of the Search service folder.
-
- 3. Select the Definitions section (shown in Defining the local search service
- at the server.)
-
- In the Definitions section of the Search service folder you set up the name
- and the service parameters of the search service.
-
- Service parameters: Setting up the service parameters for a single-workstation
- system is the same as for the server in a client/server system. Refer to
- Defining the local search service at the server. The initial values for a
- single workstation are:
-
- Max. parallel service tasks: 4
- Service tasks kept available: 3
- Time limit seconds: 3600
- Task cleanup every ms: 500.
-
-
- ΓòÉΓòÉΓòÉ 11.3.7. Verifying the SearchManager/2 installation ΓòÉΓòÉΓòÉ
-
- During installation, you installed several sample documents. Installation
- verification consists of the following steps:
-
- 1. Place sample documents in the document queue for indexing
- 2. Index the sample documents
- 3. Search for terms contained in the documents
- 4. Browse documents to see terms found.
-
- Follow the description of these steps in Getting started, beginning at A brief
- look at the indexing process.
-
-
- ΓòÉΓòÉΓòÉ 11.4. Installing dictionaries ΓòÉΓòÉΓòÉ
-
- You can install the supplied dictionaries while you install SearchManager/2,
- or, if you decide to do this later, you can install them from the
- SearchManager/2 installation diskettes. The installation package includes
- dictionaries for US English and for your local language.
-
- Dictionaries for other languages are also available (see Optional IBM licensed
- programs).
-
- To install a dictionary from the SearchManager/2 diskettes:
-
- 1. Start an OS/2 session.
-
- 2. Insert the first installation diskette into the diskette drive.
-
- 3. Make sure that the current drive is the drive containing the installation
- program; for example, drive A: if you install from drive A.
-
- 4. Type sm2inst at the OS/2 command prompt and press Enter.
-
- 5. Follow the instructions on the screen and select the appropriate language
- component. Only components not yet installed will be enabled for
- installation.
-
- Tip The dictionaries are installed in the path pointed to by the EHSNLPS
- environment variable in the CONFIG.SYS file.
-
- 6. Press Install... to start the installation.
-
- If you buy additional dictionaries: Install additional dictionaries to the
- directory pointed to by the EHSNLPS environment variable in your CONFIG.SYS
- file.
-
- If you do not install dictionaries at a client workstation: Deciding where to
- install dictionaries describes the possiblity of saving disk space in a
- client/server system by installing dictionaries only at the server
- workstation.
-
- To ensure that the client can access the file server's dictionaries:
-
- o After installing the Client/2 component, change the EHSNLPS environment
- variable in the client's CONFIG.SYS file to point to the path on which the
- dictionaries are installed, then restart your operating system.
-
- If you install dictionaries at a separate file server: Deciding where to
- install dictionaries also describes the possiblity of installing the
- dictionaries on a file server, rather than at either the client or server
- SearchManager/2 workstations.
-
- To ensure that the client and the server can access the file server's
- dictionaries:
-
- o After installing both components, change the EHSNLPS environment variable in
- the client's and in the server's CONFIG.SYS file to point to the path on
- which the dictionaries are installed.
-
-
- ΓòÉΓòÉΓòÉ 11.5. Installing SearchManager/2 without diskettes ΓòÉΓòÉΓòÉ
-
- If you have a license to install a component of SearchManager LAN Server/2 and
- Client/2 on several workstations, you may find it more convenient to do so
- without using diskettes.
-
- These are the steps:
-
- 1. Copy the diskettes to a directory on a LAN file server to which all the
- workstations have access.
-
- 2. At the workstation on which the product is to be installed, access this
- directory and enter the command
-
- SM2INST
- The installation process then runs the same way as described in Starting the
- client and server installation.
-
- You can also automate this process by using SM2INST with a response file. The
- steps are the same as above except that when you run the SM2INST program, you
- specify parameters including the name of a response file. The response file
- contains the data that you would normally enter manually if you were
- installing from diskette. This specific response file can include other
- general response files.
-
- Here is an example of an SM2INST call, showing all of its parameters. In the
- example, SearchManager/2 is being installed from a LAN drive L: to the
- workstation's D: drive. The parameters are described in SM2INST.EXE
- parameters.
-
- SM2INST options
-
- where options are:
- /A:I Install action
- /C:C:\SM2\EHSINLEN.ICF Catalog file name
- /G:L:\SM2 Path for general response files
- /L1:C:\SM2\ERROR.LOG Error log file
- /L2:C:\SM2\HISTORY.LOG History log file
- /P:"SearchManager" Product name
- /O:DRIVE Source is on disk (origin)
- /R:L:\SM2\RESPONSE.DAT Specific response file
- /S:L:\SM2 Location of source files
- /T:D:\SM2 Target directory
- /TU:C:\ Target directory to update CONFIG.SYS
- /X Process non-interactively (log errors)
-
- Here is an example of a response file, showing some of its keywords. Refer to
- Response files for a description of all the keywords.
-
- * Directory for programs
- FILE = D:/SM2
-
- * Directory for SearchManager/2 data
- WORK = D:/SM2/DATA
-
- * Directory for temporary files
- AUX1 = D:/SM2/TEMP
-
- * Directory for sample documents (SearchManager LAN Server/2 only)
- AUX2 = D:/SM2/DOCS
-
- * Component name
- COMP = IBM SearchManager Client/2
-
- * Component name
- COMP = Online documentation
-
- * Component name
- COMP = IBM SearchManager LAN Server/2
-
- * Update CONFIG.SYS automatically
- CFGUPDATE = AUTO
-
- * Overwrite files
- OVERWRITE = NO
-
-
- ΓòÉΓòÉΓòÉ 11.5.1. SM2INST.EXE parameters ΓòÉΓòÉΓòÉ
-
- These are the SM2INST command line parameters:
-
- /A:<action>
- Required. Specifies the action to be performed, where action is:
-
- D Delete
- I Install
- R Restore
- U Update
-
- If you specify this parameter, you must also specify the /P parameter.
-
-
- /C:<catalog file name>
- Required. Specifies the name and location of the catalog file that you want
- to be opened automatically. A catalog file contains information about the
- installed SearchManager/2 components.
-
- These are the catalog file names:
-
- ΓöîΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö¼ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö¼ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÉ
- Γöé LANGUAGE Γöé SEARCHMANAGER LAN SERVER/2 Γöé SEARCHMANAGER/2 Γöé
- Γöé Γöé AND CLIENT/2 Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé English Γöé "EHSINLEN.ICF" Γöé "EHSINSEN.ICF" Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé French Γöé "EHSINLFR.ICF" Γöé "EHSINSFR.ICF" Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé German Γöé "EHSINLGE.ICF" Γöé "EHSINSGE.ICF" Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Italian Γöé "EHSINLIT.ICF" Γöé "EHSINSIT.ICF" Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Spanish Γöé "EHSINLSP.ICF" Γöé "EHSINSSP.ICF" Γöé
- ΓööΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö┤ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö┤ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÿ
- A drive and directory must precede the catalog file name. Example:
-
- /C:L:\SM2\EHSINLEN.ICF
-
- /G:<include path>
- Optional. Specifies the drive and directory of the general response files
- that are included by the specific response file.
-
- Example:
-
- /G:L:\SM2
-
- /L1:<error log>
- Optional. Specifies the drive, path, and file name of the error log file.
-
- If this parameter is not specified, errors are logged to a file called
- EHSINSTS.OUT.
-
- Example:
-
- /L1:C:\SM2\ERROR.LOG
-
- /L2:<history log>
- Optional. Specifies the drive, path, and file name of the history log file.
-
- If drive and path are not specified, the drive and path where EHSINSTS.EXE
- is running is used.
-
- If this parameter is not specified, no history log is maintained.
-
- Example:
-
- /L2:C:\SM2\HISTORY.LOG
-
- /O:<source environment>
- Required. Specifies the source environment (O for origin) of the
- installation.
-
- Always set this parameter to DRIVE.
-
-
- /P:<product name>
- Specifies the name of the product.
-
- <product name> must match the value of the NAME keyword of the PACKAGE entry
- in the catalog file. See parameter /C.
-
- Example:
-
- /P:"SearchManager LAN Server/2"
-
- /R:<response file>
- Optional. Specifies the drive, path, and file name of the specific response
- file. If /R: is not specified, no response files are used.
-
- The following search order is used to find the response files.
-
- 1. The fully qualified file specification
- 2. The current directory
- 3. The file name together with the /G: invocation parameter
- 4. Each directory in the PATH environment variable
- 5. Each directory in the DPATH environment variable.
-
- Example:
-
- /R:L:\SM2\RESPONSE.DAT
-
- /S:<source location>
- Optional. Specifies the drive and path that contain the source files to be
- installed.
-
- Example:
-
- /S:L:\SM2
-
- /T:<target install directory>
- Optional. Specifies the drive and path that the product files are to be
- installed into.
-
- This parameter determines the directory to be used for installing the
- SearchManager program files. If you specify this parameter, the target
- install directory specified in the FILE paths of the response files and the
- package files is overridden.
-
- Example:
-
- /T:C:\SM2
-
- /TU:<update target CONFIG.SYS directory>
- Optional. Specifies the drive and path of the target CONFIG.SYS to be
- updated.
-
- If you do not specify this parameter, the CONFIG.SYS file on the "boot"
- drive is updated.
-
- Example:
-
- /TU:C:\
-
- /X
- Optional. Specifies that the action is noninteractive.
-
- No progress indicator is shown. Errors caused, for example, by missing
- installation information, are logged in the EHSINSTS.OUT file. Specify the
- location of this error log file using the /L1 parameter.
-
- If you do not specify this parameter, a progress indicator is shown and
- error messages are displayed. You are prompted for any information needed to
- complete the action.
-
-
- ΓòÉΓòÉΓòÉ 11.5.2. Response files ΓòÉΓòÉΓòÉ
-
- Response files are used to provide data that would normally be entered manually
- when installing from diskette.
-
- You can create several response files to cover various installation
- requirments. Each of these specific response files can be made to include
- general response files containing common installation data.
-
- Each line in a response file has a maximum line length of 255 bytes.
- Keyword-value pairs used in a response file can be in any order. Keywords are
- not case-sensitive.
-
- Here is list of the response file keywords:
-
- AUX1
- Required. Specifies an auxiliary path for installation, in this case
- for temporary files. The path for the SearchManager/2 program files is
- specified using the FILE keyword.
-
-
- AUX2
- Required. Specifies the path for sample documents (SearchManager LAN
- Server/2 only).
-
-
- CFGUPDATE
- Required. Specifies whether the CONFIG.SYS file is automatically
- updated. Values for this keyword are:
-
- AUTO: Automatically updates CONFIG.SYS
-
- MANUAL: Does not update CONFIG.SYS
-
-
- COMP
- Required. Specifies the unique name of a component of the product for
- which passed information applies.
-
- Note: Do not use quotes around the component name.
-
-
- DELETEBACKUP
- Required. Specifies whether to delete only the backup versions of the
- product or to delete the entire product. If you do not have a backup
- version, specify NO. If you have a backup version, specify YES.
-
-
- FILE
- Required. Provides the path for the directory in which the
- SearchManager/2 programs are to be installed.
-
-
- INCLUDE
- Specifies which general response files to include with a specific
- response file. The format of this keyword is:
-
- INCLUDE = filespec
-
- where filespec is the path and file name of the general response file
- to be included. If the file specification contains any global
- characters (* or ?), the first file found that matches the
- specification is included.
-
- The following search order is used to find the general response files
- specified.
-
- 1. The fully qualified file specification, if specified with filespec
- 2. The current directory
- 3. The file name together with the /G: invocation parameter
- 4. Each directory in the PATH environment variable
- 5. Each directory in the DPATH environment variable.
-
-
- OVERWRITE
- Specifies whether to automatically overwrite files during installation.
- Values for this keyword are YES and NO.
-
-
- SAVEBACKUP
- Specifies whether to save a backup version of the product when it is
- updated. Valid values for this keyword are YES and NO.
-
-
- WORK
- Required. Specifies the path for the data directory for
- SearchManager/2's data files.
-
-
- ΓòÉΓòÉΓòÉ 11.6. Deleting components of SearchManager/2 ΓòÉΓòÉΓòÉ
-
- This section describes how to remove components of SearchManager/2 from your
- workstation. You are strongly recommended to use the procedure described here
- rather than trying to delete the product files manually.
-
- Tip Saving index files
-
- When you remove SearchManager LAN Server/2 and Client/2, or the
- SearchManager/2 single-workstation system, the document index files are
- also deleted. If you want to keep them, save all files having the
- extensions DAT and IDX by copying them to a different location. These
- files are located in the directory that is pointed to by the EHSDATA
- environment variable in your CONFIG.SYS.
-
- After reinstalling, copy the saved index files back to the EHSDATA
- directory.
-
- 1. Double-click on the Installation utility object in the SearchManager/2
- window.
-
- 2. Select Delete from the Action menu in the menu bar.
-
- 3. Select the component(s) to be deleted.
-
- 4. Click on the Delete pushbutton.
-
- 5. Follow the instructions on the screen until the message Delete successful
- is displayed.
-
- 6. Having deleted the component(s), you may have to restart your
- workstation's operating system to automatically remove files that were
- previously in use.
-
-
- ΓòÉΓòÉΓòÉ 11.7. Moving index files ΓòÉΓòÉΓòÉ
-
- Sometimes an index grows to the point where it must be moved to a different
- disk. This section describes how to move the index files. The EHSDATA
- environment variable in your CONFIG.SYS file points to the directory where the
- index files are currently installed.
-
- o Create a directory on the disk to which you intend to move the index files.
-
- o The index files are located in the directory pointed to by the EHSDATA
- environment variable in your CONFIG.SYS file. Move the entire contents of
- this directory to the new directory.
-
- o Change the EHSDATA environment variable to point to the new directory.
-
- o Restart your operating system.
-
-
- ΓòÉΓòÉΓòÉ 12. Linguistic processing ΓòÉΓòÉΓòÉ
-
- SearchManager/2 offers linguistic processing in the four major areas of
- retrieval:
-
- o Indexing. When the search service analyzes documents to extract the terms to
- be stored in the document index.
-
- o Retrieval. When the search service searches through the document index to
- find the names of documents that contain occurrences of the search terms you
- have specified.
-
- o Specifying a query. When you specify the search terms for a query.
-
- o Browsing. When you browse a document that has been found after a search.
-
- In indexing, the document text is processed linguistically to extract the
- right terms for the index. This is done to make retrieval as simple and as
- fast as possible.
-
- In retrieval, the query terms are also processed linguistically to match them
- with the indexed terms.
-
- The key to successful retrieval is to use the same linguistic processing for
- retrieval as is used in indexing.
-
-
- ΓòÉΓòÉΓòÉ 12.1. Linguistic processing for indexing ΓòÉΓòÉΓòÉ
-
- When the search service indexes documents, that is, extracts terms to be stored
- in the document index, it makes a linguistic analysis of the documents.
-
- The linguistic processing that SearchManager/2 uses for indexing documents
- consists of:
-
- o Basic text analysis
-
- - Recognizing words and other text units
- - Recognizing terms that contain non-alphanumeric characters
- - Changing terms to a standard form (normalization)
- - Sentence recognition
-
- o Reducing terms to their base form
- o Part-of-speech filtering
- o Stop-word filtering
- o Splitting compound terms
- o Sentence-begin processing
- o Dehyphenation.
-
- Here is a summary of how terms are indexed when the index type is linguistic.
-
- ΓöîΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö¼ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö¼ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÉ
- Γöé DOCUMENT TEXT Γöé TERM IN INDEX Γöé EXPLANATION Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Mouse Γöé mouse Γöé Normalization Γöé
- │ KДfer │ kaefer │ │
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé mice Γöé mouse Γöé Lemmatization Γöé
- Γöé swum Γöé swim Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé computer-system Γöé computer-system, Γöé Decomposition Γöé
- Γöé Γöé computer, system Γöé Γöé
- Γöé Wetterbericht Γöé wetterbericht, Γöé Γöé
- Γöé Γöé wetter, bericht Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé a report about all Γöé report, animal Γöé Stop-word filtering: stop Γöé
- Γöé animals Γöé Γöé words A, ABOUT, ALL Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé I drive my car quickly Γöé drive, car Γöé Part-of-speech filtering: Γöé
- Γöé Γöé Γöé only nouns, verbs, Γöé
- Γöé Γöé Γöé adjectives Γöé
- ΓööΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö┤ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö┤ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÿ
-
- Figure: Term extraction for linguistic index
-
- By comparison, here is a summary of how terms are indexed when the index type
- is Precise.
-
- ΓöîΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö¼ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö¼ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÉ
- Γöé DOCUMENT TEXT Γöé TERM IN INDEX Γöé EXPLANATION Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Mouse Γöé Mouse Γöé No normalization Γöé
- │ KДfer │ KДfer │ │
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé mice Γöé mice Γöé No lemmatization Γöé
- Γöé swum Γöé swum Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé computer-system Γöé computer-system Γöé No decomposition Γöé
- Γöé Wetterbericht Γöé Wetterbericht Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé a report about all animalsΓöé report, animal Γöé Stop-word filtering: Γöé
- Γöé Γöé Γöé stop words A, Γöé
- Γöé Γöé Γöé ABOUT, ALL Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé I drive my car quickly Γöé drive, car Γöé No part-of-speech Γöé
- Γöé Γöé quickly Γöé filtering Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Sage mir Γöé Sage, sage Γöé Sentence-begin Γöé
- Γöé Γöé Γöé processing Γöé
- ΓööΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö┤ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö┤ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÿ
-
- Figure: Term extraction for precise index
-
-
- ΓòÉΓòÉΓòÉ 12.1.1. Basic text analysis ΓòÉΓòÉΓòÉ
-
- The search service processes these linguistic functions without the need of a
- dictionary.
- Recognizing terms that contain non-alphanumeric characters
-
- When indexing documents, SearchManager/2 recognizes terms even though they
- contain non-alphanumeric characters, for example:
-
- don't $14,225.23 mother-in-law 10/22/90
-
- SearchManager/2 regards the following as part of a term:
-
- o Accents and apostrophes
- o Currency signs
- o Number separator characters (like "/").
-
- SearchManager/2 also uses language-specific rules to recognize terms
- containing:
-
- o Accented prefixes in Roman languages, such as l'aventure in French
-
- o National formats for dates, time, and numbers
-
- o Contracted forms and alternatives, such as mission/responsibility, indicated
- in English using the "/" character
-
- o Trailing apostrophes in Italian words like securita'. It is usual in typed
- Italian text, when the character set does not include characters with
- accents, to type the accent after the character; for example, ╨░ is typed a'.
-
- Changing terms to a standard form (normalization)
-
- If you specify computer in a query, you will not find a document that contains
- Computer unless the search service does some normalization. In this case, the
- effect of normalization is that you find the term in the index, regardless of
- how you capitalize the word.
-
- Here is an example showing how normalization handles accents: after
- normalization, a query containing the word Вcole retrieves a document
- containing the word Ecole.
-
- If the search term includes global ("wild-card") characters, normalization is
- done before the global characters are processed. Example: BБr?er becomes
- buer?er.
-
-
- Recognizing sentences
-
- You can search for terms that occur in the same sentence. To make this
- possible, SearchManager/2 analyzes each document during indexing to find out
- where each sentence begins and ends. SearchManager/2 looks for a period
- followed by a blank character.
-
- Since a period may be part of an abbreviation, however, SearchManager/2
- distinguishes between sentences and abbreviations.
-
-
- ΓòÉΓòÉΓòÉ 12.1.2. Reducing terms to their base form (lemmatization) ΓòÉΓòÉΓòÉ
-
- If you are looking for documents that contain the term mouse, for example, then
- it is very likely that you also would like to find occurrences of the term
- mice.
-
- SearchManager/2 can reduce terms to their base form for indexing; the term mice
- would be indexed as mouse. Later, when you use the search term mouse, the
- document is found. The document is also found if you search for mice.
-
- The effect is that you find documents containing information about mice,
- regardless of which variation of the term mouse occurs in the document, or is
- used as a search term.
-
- In the same way SearchManager/2 can reduce conjugated verbs to their
- infinitive; bought, for example, becomes buy.
-
-
- ΓòÉΓòÉΓòÉ 12.1.3. Part-of-speech filtering ΓòÉΓòÉΓòÉ
-
- When SearchManager/2 extracts terms from documents to place them in the
- document index, it extracts only nouns, verbs, and adjectives; all other parts
- of speech, such as prepositions and pronouns, are filtered out. This form of
- filtering can significantly reduce the size of the index.
-
- SearchManager/2 uses the dictionary information to determine the part of speech
- of each term; it does not do any semantic analysis.
-
-
- ΓòÉΓòÉΓòÉ 12.1.4. Stop-word filtering ΓòÉΓòÉΓòÉ
-
- Stop words are, for example, prepositions and pronouns, words that occur very
- frequently in documents. Such words that the administrator considers to be
- irrelevant as search terms can be excluded from indexing and search by adding
- them to the stop-word list.
-
- Stop words must not be multiword terms.
-
- Only the exact form as contained in the stop-word list is excluded; other
- spellings or other variants are not. Stop-word filtering is in effect only when
- the stop-word list for the language concerned is used.
-
-
- ΓòÉΓòÉΓòÉ 12.1.5. Splitting compound terms (decomposition) ΓòÉΓòÉΓòÉ
-
- Germanic languages, such as German or Dutch, are rich in compound terms, like
- Versandetiketten, which means address labels. SearchManager/2 can split such
- compound terms into their components.
-
- Compound terms are always split during indexing. Compound terms are split
- during searching only on request - the user selects the Any sequence option.
-
- The components are found if they occur in any sequence in a document as long as
- they are contained within one sentence. For example, when searching for the
- German word Wetterbericht (weather report), a document containing the phrase
- Bericht Бber das Wetter (report about the weather) would also be found.
-
- SearchManager/2 tries to split a term if:
-
- o The term's language uses compound terms
-
- o The term has a certain minimum length
-
- o The term is not itself an entry in the dictionary - compounds that are
- commonly used like the German word GeschДftsbericht (business report) are in
- the dictionary.
-
- If a split is found to be possible, the term's component parts are then
- reduced to their base form.
-
- Here are some examples from Danish, German, and Dutch.
-
- Compound term Component parts
- bЫrsmСglerselskab bЫrsmСglerselskab
- bЫrs
- mСgler
- selskab
- Gewinnstreben Gewinnstrebe
- Gewinn
- Streben
- Strebe
- probleemkinderen probleemkind
- probleem
- kind
-
-
- ΓòÉΓòÉΓòÉ 12.1.6. Sentence-begin processing ΓòÉΓòÉΓòÉ
-
- In most languages, the first word in a sentence is written with a leading
- uppercase letter. In languages like German, however, each noun begins with an
- uppercase letter. So when SearchManager/2 uses the case of a term to determine
- the part of speech, the result could be incorrect for the first word in a
- sentence.
-
- To avoid this, SearchManager/2 considers both spellings (uppercase and
- lowercase) for the first word in a sentence.
-
- For example, for a German sentence beginning with Sage, both forms Sage
- (reduced to the noun Sage) and sage (reduced to the verb sagen) are then
- considered.
-
-
- ΓòÉΓòÉΓòÉ 12.1.7. Dehyphenation ΓòÉΓòÉΓòÉ
-
- During the indexing process, dehyphenation is used to normalize words
- containing a hyphen.
-
- If a hyphen has been explicitly typed in a document, dehyphenation does not
- remove the hyphen from the word during indexing. For example:
-
- co-ordination
-
- is indexed exactly as it is typed.
-
- However, if the word-processor inserts a hyphen, the hyphen is removed. For
- example, if text was formatted by the word processor like this:
-
- tainer...
-
- then the word con- tainer would be normalized to container.
-
-
- ΓòÉΓòÉΓòÉ 12.2. Linguistic processing for retrieval ΓòÉΓòÉΓòÉ
-
- When searching for documents, that is, looking in the document index for search
- terms that were specified in a query, the search service makes a linguistic
- analysis of the query terms. In doing this, the search service uses the same
- linguistic processing that it uses when indexing documents described in
- Linguistic processing for indexing. This ensures that terms are indexed and
- searched for in the same way.
-
- The linguistic processing that SearchManager/2 uses for retrieval consists of:
-
- o Basic text analysis
-
- - Recognizing terms that contain non-alphanumeric characters
- - Changing terms to a standard form (normalization)
-
- o Reducing terms to their base form
- o Splitting compound terms
- o Stop-word filtering
- o Part-of-speech filtering.
-
- Those that are not listed here do not apply to retrieval:
-
- o Sentence recognition
- o Sentence-begin processing
- o Dehyphenation.
-
-
- ΓòÉΓòÉΓòÉ 12.3. Linguistic processing when specifying a query ΓòÉΓòÉΓòÉ
-
- This section describes the linguistic processing that assists you when you are
- specifying the search terms for a query.
-
-
- ΓòÉΓòÉΓòÉ 12.3.1. Offering synonyms to enrich a query ΓòÉΓòÉΓòÉ
-
- SearchManager/2 can display synonyms for search terms and phrases.
-
- You can choose which synonyms you want to use with your search term. For
- example, if you specify combine as a search term, and select the synonyms mix
- and join, SearchManager/2 searches for documents containing combine or mix or
- join.
-
- For the English term combine the synonyms provided are:
- admix affiliate ally amalgamate associate band bind blend bloc cartel coalesce
- coalition combination combine commingle commix compound concrete conjoin
- conjugate connect couple embody faction fuse mingle mix incorporate party
- integrate pool intermingle relate intermix ring stir join syndicate league
- trust link unite marry wed meld yoke merge.
-
- If you have installed dictionaries for several languages, the list of synonyms
- contains terms in the language you have chosen.
-
-
- ΓòÉΓòÉΓòÉ 12.3.2. Offering thesaurus terms to enrich a query ΓòÉΓòÉΓòÉ
-
- If Thesaurus End User System Toolkit/2 has been installed, you can choose terms
- from a thesaurus to add to your query. Each term that you select from the
- thesaurus becomes an additional search term that you can link as you would any
- other search term using operators, like OR, AND, and NOT. This is in contrast
- to synonyms, which are associated with a particular search term (using an
- assumed OR operator).
-
-
- ΓòÉΓòÉΓòÉ 12.4. Linguistic processing for browsing ΓòÉΓòÉΓòÉ
-
- This section describes the linguistic processing that is used when you browse
- documents that have been found after a search.
-
- Linguistic analysis in the SearchManager/2 document browser is done in two
- stages:
-
- 1. Basic text analysis, normalization and term expansion.
-
- 2. Extended matching.
-
-
- ΓòÉΓòÉΓòÉ 12.4.1. Basic text analysis, normalization and term expansion ΓòÉΓòÉΓòÉ
-
- The first stage is done without accessing a dictionary. This is basic text
- analysis and normalization as described in Basic text analysis.
-
- SearchManager/2 can reduce search terms to their base form before the search
- begins. A document's terms can be reduced to their base form before being added
- to the document index. Documents are therefore found on the basis of a term's
- base form.
-
- When you browse a found document, however, you expect to see all variants of
- the base form highlighted. To highlight these variants, SearchManager/2 has to
- expand the found base term. (This is the inverse of reducing a term to its base
- form.)
-
- SearchManager/2 can produce all variants (inflections) for each term found in
- the dictionaries. These are the inflections produced for the German word gehen
- (to go):
-
- gegangen geh gehe gehen gehend gehest gehet gehst
- ging ginge gingen gingest ginget gingst gingt geht
-
-
- ΓòÉΓòÉΓòÉ 12.4.2. Extended matching ΓòÉΓòÉΓòÉ
-
- The second stage is extended matching. There can be rare occasions when
- SearchManager/2 finds a document but basic text analysis and normalization is
- not able to highlight the found term. But extended matching finds even the most
- obscure matches; see also Using extended matching.
-
- Extended matching includes all linguistic processing that is done while
- linguistically indexing; to do this, SearchManager/2 needs a dictionary on the
- client for the language of the document to be browsed.
-
- These are the occasions when extended matching can find additional matches:
-
- o The search term includes global characters and is an inflection.
-
- SearchManager/2 handles global characters and stem reduction for the search
- term and finds the corresponding documents. Text, however, that matches the
- specified search criteria is not highlighted.
-
- Example: A document contains the inflected term swam.
-
- - During indexing this term is reduced to swim.
-
- - If the search term is swi*, the above document is found, because the stem
- reduction is swim.
-
- - SearchManager/2 highlights only those words that match the term swi*.
- This is why the inflected term swam is not highlighted.
-
- o If compound words have been indexed.
-
- When a document in a Germanic language contains a compound word and is
- indexed using a linguistic index, the document index retains the parts of
- the compound word and the compound word itself. When SearchManager/2
- searches for a part of a compound word, the documents containing the
- compound word are found, but the word is not highlighted.
-
- Example: A document contains the German word Apfelbaum ("apple tree" in
- English).
-
- - During indexing this word is reduced to apfel and baum.
-
- - When the index is searched for the term baum, the term baum and the
- document that contains it is found through the index.
-
- - SearchManager/2 highlights only words in the document that match baum,
- but finds no occurrences, since the document contains Apfelbaum, but not
- Baum.
-
- o If words are hyphenated at the end of a line.
-
- If the hyphen is inserted automatically by a word processor, SearchManager/2
- can find and highlight the hyphenated word. If, however, the hyphen is typed
- by the user, the documents containing the word are found, but the word is
- not highlighted.
-
- Example: A document contains the hyphenated word container, broken at the
- end of a line like this:
-
- Another name for a folder is a con-
- tainer.
-
- - During indexing the word is normalized to container.
-
- - When the index is searched for the term container, the term and the
- document that contains it is found.
-
- - SearchManager/2 tries to highlight any words in the document that match
- container, but finds a match only if the hyphen in con- tainer was
- inserted by the text processor, and not typed by a user.
-
-
- ΓòÉΓòÉΓòÉ 12.5. How each language is supported ΓòÉΓòÉΓòÉ
-
- The following table shows how languages are supported linguistically by
- SearchManager/2, assuming that a dictionary for that language is installed. If
- a dictionary is not installed, basic text analysis is nevertheless carried out.
-
- Here is a list of the linguistic functions:
-
- Basic text analysis Dehyphenation
- Normalization Stop-word filter
- Lemmatization Part-of-speech filter
- Decomposition Synonyms
-
- The following languages support all or most of the linguistic functions:
-
- ΓöîΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö¼ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö¼ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÉ
- Γöé LANGUAGE Γöé DICTIONARY NAME Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Catalan Γöé CATALA Γöé (no decomposition) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Danish Γöé DANSK Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Dutch (preferred and Γöé NEDERLND / AKTUEEL Γöé Γöé
- Γöé modern spelling) Γöé Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé UK English Γöé UK Γöé (no decomposition) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé US English Γöé US Γöé (no decomposition) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé French (national) Γöé FRANCAIS Γöé (no decomposition) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé French (Canadian) Γöé CANADIEN Γöé (no decomposition) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé German (national) Γöé DEUTSCH Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé German (Swiss) Γöé DSCHWEIZ Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Italian Γöé ITALIANO Γöé (no decomposition) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Norwegian (Bokmal) Γöé NORBOK Γöé Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Norwegian (Nynorsk) Γöé NORNYN Γöé (no synonyms) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Portuguese (national) Γöé PORTUGAL Γöé (no decomposition) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Portuguese (Brazilian) Γöé BRASIL Γöé (no decomposition) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Spanish Γöé ESPANA Γöé (no decomposition) Γöé
- Γö£ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö╝ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöñ
- Γöé Swedish Γöé SVENSK Γöé Γöé
- ΓööΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö┤ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓö┤ΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÇΓöÿ
-
- Finnish (SUOMI dictionary) supports only:
-
- Basic text analysis
- Normalization
- Dehyphenation
- Stop-word filter.
-
- Icelandic (ISLENSK dictionary) supports only:
-
- Basic text analysis
- Normalization
- Decomposition
- Dehyphenation
- Stop-word filter.
-
-
- ΓòÉΓòÉΓòÉ 13. Thesaurus End User System Toolkit/2 ΓòÉΓòÉΓòÉ
-
- IBM Thesaurus End User System Toolkit/2, program number 5622-081, permits
- access to thesaurus data.
-
- A common use of a thesaurus is simply to find a synonym for a particular word.
- But Thesaurus End User System Toolkit/2 offers much more; it can help you find
- not only synonyms, but other related information, such as:
-
- o Hierachical relations:
-
- - Dog is a narrower term of animal
- - Dog is a broader term of poodle
- - Wheel is part of an automobile
- - Rocky Mountains is an instance of mountain regions
-
- o Associative relations: comp
-
- - Dog is associated with pets
-
- o User-defined relations:
- With this option the user can define and customize a thesaurus based on very
- specific needs; for example, in the car industry, an is made of relationship
- can be defined between product parts and the material they are made of.
-
- When Thesaurus End User System Toolkit/2 is installed, a Thesaurus push button
- is provided in the Query View of a query object. Using this push button, you
- can choose terms from a thesaurus to add to your query.
-
- Each term that you select from the thesaurus is copied into the Search term
- fields of your query to become an additional search term that you can link, as
- you would any other search term, using operators like OR, AND, and NOT. (This
- is in contrast to synonyms, which are associated with a particular search term
- using an assumed OR operator.)
-
- To display thesaurus terms:
-
- o In the Query View of a query object, select the Thesaurus push button
- (available only when Thesaurus End User System Toolkit/2 has been
- installed).
-
- This displays the Thesaurus window.
-
- Figure: Looking up terms in a thesaurus
-
- Using this window, you can choose a thesaurus, and look for terms in the
- thesaurus by displaying those terms having a particular:
-
- o Relation to the specified term, such as "narrow term," or
-
- o Relation property to the specified term, such as "hierarchical."
-
- From here you can display the Thesaurus Result window which offers you a list
- of terms related to the term you specify. The terms that you select are then
- inserted into the Search terms fields of your query.
-
- Figure: Terms found in a thesaurus
-
-
- ΓòÉΓòÉΓòÉ 14. Messages displayed by the SM2 command ΓòÉΓòÉΓòÉ
-
- When using the SM2 command you may receive a message that is identified by a
- message number. This appendix gives you help information for these messages.
-
- If you need further details refer to Troubleshooting.
-
- EHS0100 An error was detected when starting the search service. The
- EHSSVDEM.EXE file was not found.
-
- Explanation: It is possible that the path to this file is incorrect,
- or that the file has been deleted or renamed. Check that SET PATH in
- your CONFIG.SYS file points to the path that contains the file.
-
- EHS0101 An error was detected when starting the search service. EHSSVDEM.EXE
- could not be started.
-
- Explanation: It is possible that the DLL files called by this
- program could not be found, or that the EHSSVDEM.EXE file is faulty.
- Check that the LIBPATH statement in your CONFIG.SYS file points to
- the path that contains the SearchManager/2 DLLs.
-
- EHS0150 An error was detected when starting or stopping the search service.
- Stop the search service and start it again.
-
- Explanation: This can happen if there is a problem with the search
- service program at the server workstation. Try stopping the search
- service from the pop-up menu of the Search service object. Repeat if
- this seems to have no effect. If the problem persists, restart your
- operating system.
-
- EHS0200 The search service cannot be started. Check the path pointed to by
- the environment variable EHSDATA in the CONFIG.SYS file.
-
- Explanation: There are three SearchManager/2 environment variables
- in your CONFIG.SYS file:
-
- EHSDATA, EHSWORK, and EHSNLPS.
-
- Check that they are all present.
-
- EHS0210 The data file EHSADMTB.DAT is not found. Check if this file exists.
- It should be in the path pointed to by the EHSDATA environment
- variable in the CONFIG.SYS file.
-
- Explanation: Either the file could not be found or it is faulty.
-
- Check if this file exists. It should be in the path pointed to by
- the EHSDATA environment variable in your CONFIG.SYS file. Check also
- that this path does not have read-only status.
-
- If the file is on a LAN server, check that the LAN drive is still
- available.
-
- If the problem persists, check whether the disk on which this file
- resides is full.
-
- EHS0300 An error was detected when stopping the search service. Reboot the
- system and start the service again.
-
- Explanation: This can happen if there is a problem with the search
- service program at the server workstation. Try stopping the search
- service from the pop-up menu of the Search service object. Repeat if
- this seems to have no effect. If the problem persists, restart your
- operating system.
-
- EHS0310 An error was detected within the search service.
-
- Explanation: This can happen if there is a problem with the search
- service program at the server workstation. Try stopping the search
- service from the pop-up menu of the Search service object. Repeat if
- this seems to have no effect. If the problem persists, restart your
- operating system.
-
- EHS0400 An error was detected when starting the search service. The
- EHSCSSRV.EXE file was not found.
-
- Explanation: It is possible that the path to this file is incorrect,
- or that the file has been deleted or renamed. Check that the PATH
- statement in your CONFIG.SYS file still points to the correct
- directory, and that the path it points to contains the file.
-
- EHS0401 An error was detected when starting the search service. EHSCSSRV.EXE
- could not be started.
-
- Explanation: It is possible that the DLL files called by this
- program could not be found, or that the EHSCSSRV.EXE file is faulty.
- Check that the LIBPATH statement in your CONFIG.SYS file points to
- the path that contains the SearchManager/2 DLLs.
-
- EHS041x An error was detected when starting the search service. Stop the
- service and start it again.
-
- Explanation:
-
- o Most likely, the name of the search service is incorrect. Open the
- Definition section of the Settings notebook of the Search service object,
- and check that the Service name field corresponds exacly to the name
- specified in the same field at the server workstation.
-
- o It is also possible that the file EHSCXCOM.CFG could not be found or is
- faulty. Check if this file exists. It should be in the path pointed to by
- the EHSDATA environment variable in your CONFIG.SYS file. If the file is
- on a LAN server, check that the LAN drive is still available.
-
- o Maybe one of the communication programs or its DLLs could not be found or
- is faulty. Check that the LIBPATH statement in your CONFIG.SYS file
- points to the path that contains the SearchManager/2 DLLs, and that the
- PATH statement points to the path that contains the SearchManager/2
- programs.
-
-
- ΓòÉΓòÉΓòÉ 15. Troubleshooting ΓòÉΓòÉΓòÉ
-
- This appendix has two main sections:
-
- o Troubleshooting by window
-
- o Troubleshooting when starting the search service
-
-
- ΓòÉΓòÉΓòÉ 15.1. Troubleshooting by window ΓòÉΓòÉΓòÉ
-
- This section is arrranged by window. Refer to the item that describes the
- window you were using when the problem occurred. It begins with a general part
- describing symptoms that are not related to one particular window.
-
-
- ΓòÉΓòÉΓòÉ 15.1.1. Any window ΓòÉΓòÉΓòÉ
-
- In a client/server system, the connection to the server workstation is lost.
- It is possible that the search service at the server workstation has
- broken the connection to this client.
-
- One of the service parameters that are set for each search service is the
- maximum time that each service task is allowed to run. This is to prevent
- unreasonably long tasks, in particular long searches, or even error
- situations, from reducing the overall system performance. The recommended
- minimum for this parameter is 1200 seconds, that is 20 minutes. After this
- time has elapsed, the search service breaks its connection to the client.
-
- Normally, you are unaware of this interruption; when you continue to work
- with this window SearchManager/2 automatically restarts the connection to
- the search service.
-
- If, however, the connection is not automatically restarted:
-
- o Close the window that was active and then open it again. SearchManager/2
- tries to reestablish the connection to the search service.
-
- o Check that the Time limit parameter has been set to a reasonable value.
- It is defined at the server workstation, in the Definition section of the
- Settings notebook for the Search service object.
-
- o Check that no other task being handled by the search service is
- preventing your add-to-queue task from being completed in a reasonable
- time.
-
-
- The system locks.
- This may occur under extreme conditions, in particular in a
- single-workstation system, or in a client/server system when you search
- using the same workstation on which the search service is installed.
-
- Restart your system and try:
-
- o Reducing the number of applications that are running simultaneously.
-
- o Reducing the number of queries that are being processed simultaneously.
-
- o Reducing the number of queries that are open simultaneously.
-
- o Reducing the amount of time that you let periodic indexing run (not the
- periodic indexing interval) without the system being shut down and
- restarted.
-
- o Checking whether a search you are making results in an extreme number of
- matches. If so, try modifying it to narrow the search.
-
-
- In a client/server system, the connection to the server is lost, and any
- sessions using that connection lock up. No message is displayed to inform you.
- This is normal if the communication type is TCP/IP.
-
- In this case, whenever the server workstation is unexpectedly no longer
- available, a message is not displayed until two hours after communication
- is interrupted.
-
- If your system communicates using OS/2 named pipes, you are informed
- almost immediately by a message when communication to the server is
- interrupted.
-
-
- In a client/server system, you are experiencing communication problems that
- need the attention of an IBM service representative.
- In this case, the solution comes from IBM, but you can assist by providing
- trace information when you report the problem. This trace information is
- in two files:
-
- o servicename.LOG is the server trace file
- o EHSCCTRC.LOG is the client trace file.
-
- These files are in the path pointed to by the EHSWORK environment variable
- in your CONFIG.SYS file.
-
-
- In a client/server system, when you open a SearchManager/2 window after
- restarting your operating system, the message INI file not found is
- displayed.
- It is possible that you no longer have access to the disk on which the
- SearchManager/2 data files are installed. Check which disks are available
- to you on the LAN and, if necessary, restore the connection to the
- required disk.
-
-
- ΓòÉΓòÉΓòÉ 15.1.2. Add to Queue window ΓòÉΓòÉΓòÉ
-
- In a client/server system, your local drives are not displayed in the Groups
- list.
- This is normal. Only those drives are displayed to which the search
- service has access. If you were to try to place any of your local
- documents in the queue, the search service would not be able to access
- them for indexing.
-
-
- In a client/server system, you are trying to select a drive that belongs to a
- server workstation that is currently not active. It takes time before the
- system informs you.
- This is normal. The LAN requester tries several times to get a response
- from the server before informing you that the server is not active.
-
-
- ΓòÉΓòÉΓòÉ 15.1.3. Browser window ΓòÉΓòÉΓòÉ
-
- It takes time before the document text is displayed.
- If the document contains a paragraph that is particularly large, it is
- normal for the display of the document to be delayed to some extent.
-
-
- You are using a monochrome monitor, and are having difficulty seeing
- highlighted text like matches.
- Change the color settings through the Options choice in the menu bar.
-
-
- ΓòÉΓòÉΓòÉ 15.1.4. Query View window - Defaults section ΓòÉΓòÉΓòÉ
-
- In a client/server system, you are trying to work with document groups that
- are stored on a server workstation that is currently not active. It takes time
- before the system informs you.
- This is normal. The LAN requester tries several times to get a response
- from the server before informing you that the server is not active.
-
-
- In a client/server system, your local document groups are not displayed in the
- Restrict to document groups list.
- This is normal. You can restrict your search only to those document groups
- to which the search service has access.
-
-
- ΓòÉΓòÉΓòÉ 15.1.5. SearchManager folder ΓòÉΓòÉΓòÉ
-
- When you open the SearchManager folder, the Search service object is not
- visible.
- It is possible that the Search service object is scrolled out of sight.
-
- To correct this:
-
- 1. Click the right-hand mouse button anywhere within the SearchManager
- open folder to display the folder's pop-up menu.
-
- 2. Select Arrange. If you are working with a client/server system, it is
- possible that the window needs to be refreshed. To do this:
-
- 1. Click the right-hand mouse button anywhere within the SearchManager
- open folder to display the folder's pop-up menu.
-
- 2. Select Refresh.
-
-
- ΓòÉΓòÉΓòÉ 15.2. Troubleshooting when starting a search service ΓòÉΓòÉΓòÉ
-
- If an error occurs when starting a search service, try to solve it using the
- information displayed by the SM2 command (see Messages displayed by the SM2
- command), or by using the information that is displayed after selecting Start
- on the pop-up menu of the Search service folder.
-
- If these methods do not lead to a solution, you can display more detailed
- diagnosis information by entering the following command in an OS/2 session at
- the workstation where the search service is installed:
-
- o EHSCSSRV servicename
-
- where servicename is the name specified in the Definitions page of the Search
- service object.
-
- This command may display the following information:
-
- 1303 TCP_INSTALL_PROBLEM
- Explanation: You are using TCP/IP, but the TCP/IP installation is
- faulty. Check that TCP/IP is installed correctly.
-
-
- 1307 SERVICE_IS_REMOTE
- Explanation: You tried to start a search service located on a
- different machine.
-
-
- 1325 INV_PARMS_ERROR
- Explanation: The command has not been entered correctly.
-
- 1305 INIT_FAILED
- 1312 DOSEXECPGM_ERROR
- 1316 CFG_FILE_OPEN_ERROR
- 1318 CFG_FILE_READ_ERROR
- 1319 CFG_FILE_CLOSE_ERROR
- 1320 DOSLOAD_COM_METHOD_ERROR
- 1323 OPENTRACE_ERROR
- 1350 CFG_SERVER_NOT_FOUND_ERROR
- 1398 INV_COMTYPE_ERROR
- Explanation: These problems may be due to your environment; it is
- possible that there is an error in your CONFIG.SYS file. Check the
- following points. If the problem persists, restart your operating
- system. If SearchManager/2 files are faulty or missing, reinstall
- SearchManager/2. Remember to save your index files before
- reinstalling SearchManager/2; refer to Deleting components of
- SearchManager/2.
-
- Check that the PATH statement points to the directory that contains
- the SearchManager/2 programs.
-
- Check that the LIBPATH statement points to the directory that
- contains the SearchManager/2 DLLs.
-
- Check that the EHSDATA environment variable points to the directory
- that contains the SearchManager/2 data file EHSCXCOM.CFG.
-
- Check that the EHSDATA environment variable points to a directory
- that has sufficient free space available. If not, you should move
- your document index files to a different disk. Refer to Moving index
- files.
-
- Check that the EHSWORK environment variable points to a directory
- that has sufficient free space available.
-
- If the EHSDATA or EHSWORK environment variable in your CONFIG.SYS
- file points to a directory on a LAN file server, make sure that the
- file server is available and that you have write access to the
- directory.
-
- 1301 MAX_FILE_HANDLE_ERROR
- 1306 BEGTHREAD_ERROR
- 1308 DOSPOSTSEM_ERROR
- 1310 DOSKILLPROC_ERROR
- 1311 DOSWAITCHILD_ERROR
- 1314 CONNECTOR_DOSPOSTSEM_ERROR
- 1321 DOSQRY_PROCADR_ERROR
- 1322 SHARED_MEM_ERROR
- 1326 CREATEMEMSEM_ERROR
- 1327 CREATETASKSEM_ERROR
- 1328 CREATECLIENTSEM_ERROR
- 1329 CONNECTOR_DOSCLOSESEM_ERROR
- 1330 MAIN_DOSCLOSESEM_ERROR
- 1340 CLOSECLIENTSEM_ERROR
- 1360 CONNECTOR_FREETASKSEM_ERROR
- 1370 CXCLOSE_ERROR
- 1380 CXCREATE_ERROR
- 1390 CXDISCONN_ERROR
- 1395 OPENFREETASKSEM_ERROR
- 1397 SETPRIORITY_ERROR
- 1398 INV_COMTYPE_ERROR
- 1399 RESETTASKSEM_ERROR
- Explanation: The above problems may be due to resource constraints
- on your workstation.
-
- 1. Close some applications.
-
- 2. Check that the drive that contains the SWAPPER.DAT file has sufficient
- free space available.
-
- 3. Restart your operating system.
-
-
- ΓòÉΓòÉΓòÉ 16. Some books you may need ΓòÉΓòÉΓòÉ
-
- o IBM SearchManager Client/DOS for Windows: Information Retrieval, SH12-6070
-
- o IBM SearchManager/2: Programming Interfaces, SH12-5987
-
-