Collection of Tools & Utilities

home *** CD-ROM | disk | FTP | other *** search

/ Collection of Tools & Utilities / Collection of Tools and Utilities.iso / tex / indxx802.zip / MANUAL.DOC < prev next >

Wrap

Text File | 1991-11-14 | 189KB | 4,524 lines

- INDEXX - Professional Version User's Manual Version No. 8.02 Copyright 1985-1991 by: Norman Swartz 1053 Ridley Drive Burnaby, B.C. Canada V5A 2N7 tele. (604)-420-7454 9:00-20:00 Pacific time Contents Chapter 1 Introduction 1 1.1 Notice to users updating from previous versions of INDEXX . . . . . . . . . . . . . 1 1.2 What INDEXX does . . . . . . . . . . . . . . 1 1.3 Minimum requirements . . . . . . . . . . . . 3 1.4 Maximum operating parameters . . . . . . . . 3 1.5 Files on the distribution diskette . . . . . 4 1.6 User's fee . . . . . . . . . . . . . . . . . 5 1.7 Reporting problems and suggesting improvements . . . . . . . . . . . . . . . . 5 1.8 Credits . . . . . . . . . . . . . . . . . . . 6 1.9 Disclaimer . . . . . . . . . . . . . . . . . 6 Chapter 2 Tutorial 7 2.1 Copying the distribution diskette and installing INDEXX . . . . . . . . . . . . . . 7 2.2 Obtaining a printed copy of the Manual . . . 7 2.3 Loading INDEXX . . . . . . . . . . . . . . . 8 2.4 Running INDEXX . . . . . . . . . . . . . . . 8 2.5 Creating an index file . . . . . . . . . . . 9 2.6 Entering/Editing data in your index file . . 9 2.7 Alphabetizing your index file . . . . . . . 22 2.8 Formatting and printing your index file . . 22 2.9 Exploring the files on your data disk . . . 24 2.10 Retrieving a file in progress . . . . . . 24 Chapter 3 Detailed operation of INDEXX 26 3.1 The Main Menu . . . . . . . . . . . . . . . 26 3.2 Option 0 - Changing the defaults . . . . . 26 3.2.1 Choose screen colors (full-color monitor only) . . . . . . . . . . . . 27 3.2.2 Set operating parameters . . . . . . . 27 3.2.3 Command line parameters . . . . . . . 29 3.2.4 Using a .BAT file . . . . . . . . . . 30 3.3 Option 1 - Creating an index file . . . . . 30 3.4 Option 2 - Retrieving a file in progress . 31 3.4.1 Errors detected in reading the data file . . . . . . . . . . . . . . . . . 31 3.5 Option 3 - Updating the index . . . . . . . 32 3.5.1 Normal data entry mode . . . . . . . . 32 3.5.2 String data entry mode . . . . . . . . 32 3.5.3 Subheadings . . . . . . . . . . . . . 33 3.5.4 Internal file fields . . . . . . . . . 33 3.5.5 Partial matches . . . . . . . . . . . 34 i 3.5.6 Wildcard searches . . . . . . . . . . 36 3.5.7 Typefaces: italics, boldface, etc. . . 36 3.5.8 `As-if' codes (forced sorting) . . . . 37 3.5.9 Foreign alphabets . . . . . . . . . . 37 3.5.10 Cross-referencing . . . . . . . . . . 37 3.5.11 Page citations . . . . . . . . . . . 38 3.5.11.1 Cross-references among page citations 40 3.5.12 Automode . . . . . . . . . . . . . . 41 3.5.13 Editing . . . . . . . . . . . . . . . 42 3.5.13.1 Editing the input line 43 3.5.13.2 Retrieving and editing a file record 43 3.5.13.3 Searching for records with error flags [###] 46 3.5.14 Review of function key usage . . . . 46 3.5.15 Error messages in Option 3 . . . . . 47 3.5.15.1 Message 1: "Only 2,000 entries permitted" 47 3.5.15.2 Message 2: "Out of memory" 47 3.6 Option 4 - Alphabetize index . . . . . . . 48 3.6.1 Sort-1 / Sort-2 . . . . . . . . . . . 48 3.6.2 Letter-by-letter / Word-by-word sorts . 48 3.6.3 Float / Sink . . . . . . . . . . . . . 48 3.6.4 Character 225 . . . . . . . . . . . . 49 3.6.5 File header . . . . . . . . . . . . . 49 3.6.6 Ignoring `little' words . . . . . . . 49 3.6.7 Combining of records . . . . . . . . . 50 3.6.8 Sorting page citations; Problem cases . 50 3.7 Option 5 - Formatting index for printing . 51 3.7.1 Formatting options . . . . . . . . . . 51 3.7.2 Formatted print file . . . . . . . . . 53 3.8 Option 6 - Make a backup copy of index . . 53 3.9 Option 7 - Exit to DOS . . . . . . . . . . 53 3.10 Review of INDEXX files . . . . . . . . . . 54 3.11 Editing INDEXX files with a word- processor . . . . . . . . . . . . . . . . 54 Chapter 4 Supplementary program - INDXPLUS 56 4.1 Option 8 - Merging files . . . . . . . . . 56 4.2 Option 9 - Formatting (oversized) .IND files . . . . . . . . . . . . . . . . . . . 58 4.3 Option 10 - Splitting oversized .IND files . 58 4.4 Option 11 - Converting old-style INDEXX files to current format . . . . . . . . . . 58 Chapter 5 Questions and answers 59 Appendix A Sorting algorithms 62 A.1 Pseudo-records . . . . . . . . . . . . . . 62 A.2 Special cases . . . . . . . . . . . . . . . 63 Appendix B Updating from previous versions of INDEXX 64 ii B.1 Changes in the design of record fields . . 64 B.2 Merging old .IND files . . . . . . . . . . 65 Index 66 iii Chapter 1 Introduction 1.1 Notice to users updating from previous versions of INDEXX Users who wish to use .IND files created by previous versions of INDEXX should read Appendix B (p. 64). 1.2 What INDEXX does One of the most tedious tasks facing authors, editors and researchers is constructing a book index from printed copy. To be sure, many of the better contemporary computer word-processors provide indexing facilities for manuscripts and books produced using those word-processors.[1] The shortcoming common to all such word-processor indexing programs, however, is that they can be used only on data files which exist in computer readable format, e.g. can be used only on books stored on computer disks or tapes etc. Such word-processor based indexing programs are ill-suited for indexing books which exist on paper. Thus even though one might prepare a manuscript using such a word- processor, if the publisher decides to typeset the book afresh, then the page numbers of the typeset copy will not correspond to those of the manuscript. Usually in such a case, the publisher will require the author or editor to produce a new index from the page proofs. INDEXX is designed to aid in the constructing of indexes for books which exist in printed format, e.g. for (1) new manuscripts in press, using page proofs, and for (2) published works (e.g. archival materials) which lack indexes which are needed as research tools. It serves much as did old-fashioned card systems, the difference being that the computer, not the person compiling the index, looks after searching for matches with previously entered data, alphabetizing the results, sorting page citations into ascending order, and printing the index in any of a variety of formats. ----------- 1. Indeed, the Index for this Manual was produced by FinalWord II (copyright FW Corp.) and not by INDEXX itself. - 1 - It can create an index to one inner level of `nesting', i.e. can list subheadings under a heading, e.g., musical recording analog vs. digital, 88-9, 205, 400 Dolby A (noise reduction), 16 financial investment, 3, 11, 18-20 scheduling, 22, 41-7, 212 In addition, INDEXX can sort data in a great many different ways. 1. It can sort, for example, (a) with capitalized and lowercase entries intermingled in alphabetical order, or (b) with all capitalized and all lowercase entries alphabetized separately and printed in two separate lists, the latter method (b) producing a `Names' index followed by a `Subject' index. That is, you can produce both Names and Subject indexes simultaneously. 2. For any particular entry, you can stipulate a forced `as-if' sorting, e.g. you can instruct INDEXX to sort "2nd World War" as if it were "Second World War". Or suppose you are producing both a Names and a Subject index; you could direct INDEXX to sort, e.g., "Kantian philosophy" as if it were a subject (e.g. as if it were "kantian philosophy"). 3. It can sort in letter-by-letter fashion (explained below, page 48), or in word-by-word fashion. 4. It can `float' names beginning with "Mac" or "Mc" to the top of the list of entries beginning with "M", e.g., at your option it will alphabetize "McCarthy" before "Maguire". 5. The sorting routine will ignore up to 50 user-specified `little' words, e.g. if you instruct it to ignore "the" and "on", it will alphabetize "The Mill on the Floss" as if it were "Mill Floss". And finally, the sorting procedure also handles French, German, and Spanish alphabets correctly, e.g. u-umlate (ASCII character 129) is alphabetized as if it were "u", e-grave as "e", etc. INDEXX provides two different formats (switchable `on the fly') - `normal' and `string' - for inputting data. In the `normal mode', for example, it is not necessary to re-order persons' names. One may enter names as one would normally refer to a person, e.g., one can enter "Arthur Conan Doyle", and the program will re-order the name, placing the surname first, and the given name(s) last (e.g. the output would read "Doyle, Arthur Conan"). INDEXX has a built-in editing facility. Any record, or block of records, may be retrieved and edited on-screen. In addition, - 2 - all of the files which INDEXX produces are in standard ASCII format, and thus, if you like, can be edited with any word- processor. INDEXX allows you to use up to 26 different typeface codes, e.g. for italics, boldface, underlining, small caps, etc. INDEXX is able to search for previously entered data on the basis of only a few initial letters (i.e. INDEXX searches the file [database] using a kind of so-called fuzzy logic). Thus having entered the full name "Arthur Conan Doyle", for the first occasion of its occurrence in the book, you need thereafter enter only a few letters for subsequent references to Doyle, e.g. need enter only "A Do" [or even just "Do"] and INDEXX will find and display all previous (partial) matches. You can then choose the correct match from among the `partial' (or near) matches displayed, and INDEXX will add the latest page reference(s) to the earlier one(s) stored in the file. Like the sorting facility, the formatting facility in INDEXX offers a great many options, including, for example, - variable line width (30 to 72 characters) or `infinite' line width; - specifiable line turnover; - variable vertical spacing; - compacted or separated initials on persons' names; - optional comma between heading and page citations; - aligned or run-in style for subheadings; - specifiable indent for subheadings; and - optional prefatory dash on subheadings (similar to the items in this very list). 1.3 Minimum requirements MS/IBM DOS computer 410 kilobytes (419,840 bytes) available RAM 1 disk drive DOS 2.0 or later (optional) printer (optional) word-processor capable of editing ASCII files - 3 - 1.4 Maximum operating parameters INDEXX - References (records) in an index file held in computer memory . . . . . . . . . . . . . . . . . . . . . . . 2,000 - Characters (bytes) per file . . . . . . . . . . . >150,000 - Characters (bytes) per record . . . . . . . . . . >10,000 - Separate page citations per record . . . . . . . . . . 200 - Partial matches on data entry . . . . . . . . . . . . . 50 INDXPLUS The supplementary program INDXPLUS (on the distribution diskette) can collate sorted indexes created by INDEXX, producing longer ones as follows: - References (records) in an index file . . . . . . No limit - Characters (bytes) per file . . . . . . . . Disk capacity - Characters (bytes) per record . . . . . . . . . . >10,000 - Separate page citations per record . . . . . . . . . 1,000 Using INDEXX, you could produce indexes, for example, on a chapter-by-chapter basis. Using INDXPLUS, you could then combine these several files into a complete alphabetically sorted index, even if the resulting index exceeds the memory capacity of your computer. 1.5 Files on the distribution diskette There are seven files included on the distribution diskette(s): READ.ME brief instructions on how to begin and late additions, corrections, etc., if any, to the current Manual MANUAL.BAT a batch file for printing this Manual (see page 7) MANUAL.DOC this Manual INDEXX.EXE the working copy of INDEXX INDXPLUS.EXE supplement to INDEXX (see Chapter 4) - Either - PAYMENT.DOC a payment remittance form for user-made copies of INDEXX and INDXPLUS (see Section 1.6) - 4 - PAYMENT.BAT a batch file for printing the Remittance Form - Or - ORDRFORM.DOC an Order Form for users of the demonstration version to order a full working version ORDRFORM.BAT a batch file for printing the Order Form 1.6 User's fee INDEXX is not public-domain software. If you are using a copy which you have not purchased from the author or the distributor, you are in violation of copyright law. A Remittance Form (PAYMENT.DOC) is included on the distribution diskette. The price is $99.00 US ($115.00 Canadian) for a single-user. To print the Remittance Form, place the distribution diskette in Drive A, and at the DOS prompt ">", type: PAYMENT<Enter> Users of the demonstration version who wish to order the full working copy, can obtain an Order Form by placing the distribution diskette in Drive A, and at the DOS prompt ">", typing: ORDRFORM<Enter> 1.7 Reporting problems and suggesting improvements I would be pleased to receive your comments on the use to which you've put INDEXX. And I'd like, too, to be informed of whatever errors (so-called `bugs') you find in the program. If your printout is not what you expected, perhaps you could send me a copy with a note of explanation. If, on the other hand, the program should `crash' (stop running) and an error message appears on the monitor to the effect " ... in MODULE ...", try to take a so-called screen-dump (turn on your printer and press the <PrtSc>-key while holding down the <Shift>-key), and mail me the printed copy. I will need the error message and its cryptic numbers in order to try to track down and correct the problem. I might add, however, by way of reassurance, that INDEXX is fairly well tested (debugged), and no user of any previous version has reported losing data because of an internal programming flaw in INDEXX. If you would like to suggest improvements, please do write to me. My address appears on the cover of this Manual. - 5 - 1.8 Credits The Manual and the source code for INDEXX were written using the FinalWord II word-processor manufactured by FW Corp. The source code was written in Microsoft's Quickbasic dialect of Basic, and was compiled with that company's 4.50 compiler. Certain assembly-language subroutines have been incorporated in INDEXX. These were written by Thomas Hanlin III and Wayne M. Hammerly. These routines are available in the ProBas profess- ional programming library available from Hammerly Computer Services, Inc., 8008 Sandy Spring Road, Laurel, MD 20207. 1.9 Disclaimer This program is distributed "as-is". In no event will the author or the distributor be liable for any damages, including lost data, lost profits, lost savings or other incidental or consequential damages arising out of the use of, or inability to use, INDEXX and/or the additional software provided on the distribution diskette. - 6 - Chapter 2 Tutorial This manual is rather lengthy: it is intended to be a comprehen- sive reference guide. But its sheer size may give the mistaken impression that INDEXX is difficult to use. It is not. To see how straightforward INDEXX is, let's put it through its paces even before you read through to the end. If you follow the tutorial instructions in this chapter, you will have learned most of what you need to know. The finer points, in the next chapter, can be safely left for another day after you're more experienced. Understand that this Manual is similar to manuals which accompany word-processors. That is, such manuals explain the operations of the software; but they do not instruct you how to write well or what to index. These latter (creative) skills are best taught by very different kinds of manuals. If you are new to indexing, you may want to read through the relevant sections in the Chicago Manual of Style (University of Chicago, 1982) for guidelines and pointers on how to choose effectively what you include in your index. 2.1 Copying the distribution diskette and installing INDEXX If you have not already done so, you should make a working copy of the distribution disk(s) of INDEXX, and retire the original(s) to a safe place. Make sure (each) distribution diskette has a write-protect tab on it, and then copy the files INDEXX.EXE and INDXPLUS.EXE onto a formatted diskette (e.g. B:\) or into a subdirectory (e.g. C:\INDEXX\) on your hard drive. These copied files will be your working version of INDEXX. 2.2 Obtaining a printed copy of the Manual If you are reading this Manual on the monitor and not from a printed copy, you can make a printed copy simply by placing a diskette with copies of the files MANUAL.BAT and MANUAL.DOC in drive A, and entering A:\MANUAL in response to the DOS prompt. Then just follow the instructions on the monitor for readying your printer. - 7 - 2.3 Loading INDEXX Floppy disk systems - On floppy disk systems, place your working diskette, with a copy of INDEXX.EXE in drive A, and then enter A:\INDEXX (note the double "XX") in response to the DOS prompt. Since the entire program is loaded into memory, you may remove the disk containing the program from the drive after the program has loaded. On single-drive systems, you can then place your data disk in this same drive. Hard disk systems - On hard disk systems, before you load INDEXX for the first time, create a subdirectory for your data files. This subdirectory will be used in the tutorial which follows shortly. Since DOS can create only one subdirectory per instruction, you will need two "mkdir" (MaKe DIRectory) instructions for our purposes. c: <Enter> cd c:\ <Enter> mkdir c:\book <Enter> mkdir c:\book\chapters <Enter> Now you are ready to load INDEXX. Switch to the subdirectory in which INDEXX.EXE resides. (Note the double "XX" when you type "INDEXX".) E.g. cd c:\indexx <Enter> indexx <Enter> (If INDEXX does not load, in particular if you get an error message "Insufficient memory", turn to page 59 for corrective action.) 2.4 Running INDEXX The first thing INDEXX does when loaded is test to see which sort of monitor adapter card is active in your computer. If you are using a color graphics adapter card (CGA, EGA, VGA, or Super VGA), INDEXX will pause to ask whether that adapter is connected to a Full-color monitor or to a Monochrome monitor. Choose F or M.[2] If you are using the monochromatic adapter, you cannot be using a full-color monitor and this question will be skipped. ----------- 2. In the next chapter, I'll tell you how to configure INDEXX so that this question can be skipped (see p. 29). - 8 - 2.5 Creating an index file Next in order are the opening logo and then the Main Menu. Take a moment to familiarize yourself with then Main Menu, then choose Option 1 - Begin a new index (i.e. create an index file). Notice when you choose item "1" on the Main Menu that the keys are set to `hot', i.e. you press a single key and do not have to press the `Enter'-key to get INDEXX to respond. In general, when a single-key response is requested in INDEXX, the keys are `hot'; but when you need to enter several characters, you will need to press the `Enter'-key to inform INDEXX that you have reached the end of the string of input data. Hereinafter, I'll indicate the `Enter'-key (sometimes called the Carriage Return) by: <Enter> As soon as you've pressed the "1"-key at the Main Menu, INDEXX will switch to Option 1; the monitor will clear, and you will be asked which drive and path you wish to use for your data file. If you are using a floppy disk for your data files, be sure to place a formatted, but otherwise empty, disk in the drive you are about to specify, and then enter the so-called `drive spec.' for that drive, e.g. A:\<Enter> or B:\<Enter> If you have a hard drive, INDEXX will work faster. For a hard drive, specify the subdirectory you created (see p. 8) before loading INDEXX, viz. C:\BOOK\CHAPTERS\<Enter> Whether you are using a floppy disk or a hard drive, you will next be asked for a name to be given to your new data file. Just type: TEST-ONE<Enter> After the new file has been created, pressing the <Space-Bar> will return you to the Main Menu. 2.6 Entering/Editing data in your index file You're ready to start entering data in your file. Choose Option 3 - Edit/Update (modify/add to) index. The Main Menu will be erased and replaced by the display for Option 3. At the very top of the monitor you'll see the name of the file you're currently working on (it will read "TEST-ONE.IND"). On - 9 - line four will appear the last item you entered. When you first start up, this line will read "[Nil]". And three lines below, at the far right, you can see that the file has the status EMPTY (i.e. has no data in it). As you perform each of the instructions below, you may want to check them off in the spaces provided [ (___) ], so that you can be sure to perform every one. As you enter data, that data will appear at the cursor, about halfway down the monitor. If you make a mistake typing, the backspace key, the delete key and the cursor keys can all be used to correct your data before you press <Enter>. The status of the CapsLock key is shown by an arrow on the monitor immediately to the left of the prompt ["Heading:", "Subheading:", or "Page(s):"]. To begin, make sure the keyboard is shifted to lowercase (the arrow to the left of "Heading:" should be pointing downward), and then type for the heading: apples <Enter> (___) Immediately a `query box' will appear a few lines below and will tell you that this heading doesn't match any previous heading, and will ask you to confirm that it is to be added to the file. Answer Y (for "Yes" [upper or lowercase is immaterial]): Y (___) (Note that you don't have to press <Enter>. The Y-key and N-key are set to `hot' for this query.) As soon as you reply Y, the cursor will jump down to the next line, ready for you to enter a subheading. We'll skip a subheading, by pressing <Enter>: <Enter> (___) Again the cursor will jump down a line, ready now for you to enter the page(s). 45 <Enter> (___) Immediately, this (new) item will be placed on line four as the `last item' entered. You will also note that the status of the file which was originally EMPTY has now changed to UNSORTED (i.e. the file now contains data, but it is not alphabetized). Next type the following. (Don't worry about the page numbers being out of sequence.) museums <Enter> (___) Y (___) <Enter> (___) 34, 55, 21 <Enter> (___) - 10 - This second entry will now replace "apples \ 45" on line four. It, too, will be `new'. Now type: apples <Enter> (___) No confirmation will be asked in this case, since INDEXX will find a perfect match to a heading that is already in the file. Instead, INDEXX will proceed immediately to position the cursor ready for the (optional) subheading. This time we will use a subheading. Type: sauce <Enter> (___) Y (___) 23 <Enter> (___) For your next item, (taking care to leave a space between "D." and "Hume"), type: D. Hume <Enter> (___) As expected, the query box will appear. But before you answer "Y", take careful note how INDEXX has reformatted the name. You had typed "D. Hume", but what now appears is "Hume, D.". INDEXX has split off the very last word (viz. "Hume") as the surname, and assigned whatever was previous to it (viz. "D.") as given name(s). Confirm that this is correct: Y (___) Skip over the subheading, and enter the page(s): <Enter> (___) 67 <Enter> (___) Now, let's try this again, but with a longer name, and with a subheading. Type: Charles Philip Arthur George Windsor <Enter> (___) Observe how INDEXX split this entry. Again confirm that this is correct, and then enter the subheading and the page(s). Type: Y (___) education <Enter> (___) Y (___) 75 <Enter> (___) When Prince Charles' lengthy name is placed on line four of the monitor as the `last item' entered, there isn't room for the entire name: it will be shortened to fit. Don't worry, however. All of the heading has been recorded in your file. To see this, you can use function key <F2>. Doing so, you will then be asked whether you want to view the file `as-is', or whether you want to `rough-sort' it first. Choose "V", for `view as-is': - 11 - <F2> (___) V (___) Look over the file a bit. (The complete entry for Charles occurs as the last record in the file.) You will see at the outset a twelve-record header which contains information that various Options require about your file. Beginning with the 13th record, you can see that each RECORD consists of three FIELDS - the heading, the subheading, and the page citation(s). Each field is marked out with double quotation marks and the three fields are separated from one another by commas. INDEXX recognizes SUBFIELDS only within the third field, i.e. among the page citations; subfields are demarcated by commas. If there happens to be no subheading under a given heading - e.g. in the case of "museums" - the second field will simply be null (vacant). Carriage-returns at the end of each record are shown on the monitor as half-diamonds. (These carriage returns will likely be represented by some other character, or may even be invisible, if you edit the file in your word-processor.) What you are looking at is `raw data'. To be sure, it doesn't much resemble a polished book index. The headings are not alphabetized, the page citations are out of order (e.g. on "museums"), and the format bears little resemblance to the neat columns of familiar indexes. It is the job of Option 4 to alphabetize and sort this raw data, and then of Option 5 to format it in standard form. When you're ready, press <F10> (___) to return you to where you left off in Option 3. It's time to try the Automode. The Automode feature has INDEXX supply the page number without your having to type it. To turn on the Automode, press function key <F3> (___). About a quarter of the way down the monitor, a line will appear telling you that the Automode is ready to enter data for page 75 (i.e. continuing on the last page you specified). Type: apples <Enter> (___) pies <Enter> (___) Y (___) For the page(s), do not type any number at all, but simply press <Enter>: <Enter> (___) You will see that "apples \ pies" now appears on line four bearing the page citation "75". You can change the setting of the Automode by using function keys <F5> and <F7>. Press <F5> twice: - 12 - <F5> (___) <F5> (___) The Automode will update to page 77. Now type: apples <Enter> (___) <Enter> (___) <Enter> (___) Line four will now show "apples" recorded at page 77. You can also update the Automode simply by entering an updated page number with your data, i.e. by entering a page number as if the Automode were not on. Type: art <Enter> (___) Y (___) sculpture <Enter> (___) Y (___) 85-93 <Enter> (___) The Automode has updated to page 85 (not to 85-93). Now type, without using a page number: apples <Enter> (___) varieties <Enter> (___) Y (___) <Enter> (___) A citation for the subheading "varieties", under the heading "apples", is recorded for page 85. You can see why the Automode is programmed to use only the first number in a sequence of page citations that might be entered on any given item, e.g. why it did not use all of the sequence "85-93" entered a moment ago for "art \ sculpture": it's very unlikely that subsequent data entries would be for precisely the same span (range) of pages. Having seen how the Automode works, we'll turn it off now by pressing <F3>: <F3> (___) We are coming up to the 10th entry. As soon as the 10th entry is added to the file in memory, the entire file will be automatically written to disk. (If you are using a hard drive, this automatic saving-to-disk may occur so quickly that you are not aware of it.) Your data are written out to disk after every ten entries so as to insure that little will be lost in case of a power failure, etc. For your 10th entry, let's enter another page citation for Prince Charles. Only this time we won't have to type his full name. INDEXX's ability to retrieve partial matches makes the entry of successive page citations on previously entered headings - 13 - and subheadings remarkably easy and assures identity of spelling on each occasion. We need type only a few letters and INDEXX will find a match on the previously entered record. Suppose you want to add page "235" to the earlier subheading "education" under "Windsor, Charles Philip Arthur George". Typing all of this out in full would be a bit daunting. But you need, in fact, type remarkably little. Begin by retrieving the heading "Windsor, Charles ... etc.", by typing: W <Enter> (___) The monitor will clear, and you'll now be presented with a more elaborate query box listing all the `partial matches' INDEXX has found in the data file. In this case, there's only one. Confirm the match: <Enter> (___) You're ready now to retrieve the subheading "education" and to add the page(s): e <Enter> (___) <Enter> (___) 235 <Enter> (___) That was pretty easy, you'll have to admit. Not many keystrokes used, were there? As soon as you finish entering this 10th item, INDEXX will disable your keyboard and flash a message that it is writing your data file to disk. When the writing is completed, you may resume entering data. For the 11th entry, type: David H <Enter> (___) Again, a list of partial matches will appear. Confirm that this is a match with the earlier entered "Hume, D.". Respond: <Enter> (___) Yet another query box will appear, one you've not seen earlier. This latest entry "David H" contains not only an abbreviation ("H") for the earlier surname ("Hume"); it also contains a given name ("David") which is longer than the one ("D.") presently in your data file. INDEXX has taken the longer surname from the earlier entry and combined it with the longer given name from this latest entry and displayed the result as the "proposed revision". (Pretty clever, wouldn't you say?) Accept the revision: Y (___) - 14 - The monitor will revert to the familiar display for Option 3, and you can see that the input beside "Heading:" now reads "Hume, David". Skip the subheading, and add the page(s): <Enter> (___) 49 <Enter> (___) Let's now see what happens if you violate the proper format for a data item. Type the following improper entry, where you add a dangling comma after "apples": apples, <Enter> (___) An error message has appeared telling you that you have used a comma incorrectly. Press the <Space-Bar> (___) to remove the error message, and then correct the mistaken entry. (Here you can conveniently use the <Backspace> key to erase the comma.) Skip the subheading, and then enter an improper set of page numbers, specifically: <Backspace> <Enter> (___) <Enter> (___) 56, , 7 <Enter> (___) Again, you'll be told that you've misused commas. Again press the <Space-Bar> (___) to remove the error message. But this time, let's erase the entire mistaken line. To do this, press the <Esc> (escape) key. <Esc> (___) 56, 7 <Enter> (___) (If you want to erase not just one field - i.e. the current heading, subheading, or page(s) - but all the input data displayed on the monitor, use <F6>. [See the `info' line at the very bottom of the monitor.]) The `normal' data entry mode (which we've been using) is suitable for single-word entries (e.g. "apples") and persons' names (e.g. "David Hume"). It always reformats the entry by switching the very last word into the leading position. Thus it would reformat "New Jersey" (incorrectly) as "Jersey, New", and "Arthur Sullivan and William Gilbert" (incorrectly) as "Gilbert, Arthur Sullivan and William". To enter data without reformatting, you must use the alternative, `string', mode for data entry. To switch back and forth between `normal' and `string' modes, for either headings or subheadings, press the <F1> function key.[3] Let's try this for the last example just ----------- 3. No reformatting is available (or wanted) for page entries, i.e. only the string mode is available for page entries. - 15 - mentioned. Begin by switching from the `normal' to the `string' mode. <F1> (___) Immediately the reminder on line nine, in the middle of the monitor, will change. Now type: Sullivan, Arthur, and William Gilbert <Enter> (___) Y (___) <Enter> (___) 67 n.2, 23, 100 <Enter> (___) On completing this latest entry, you are returned to the Normal data input mode. The Normal mode is the default setting for headings; the string mode is the default setting for subheadings. (You can change the default for the heading by the use of the <F9> function key, but we'll leave that for you to learn in the next chapter.) Let's sum up a bit. As you have seen, whenever you enter a heading (e.g. "H" for "Hume") or a subheading (e.g. "e" for "education"), INDEXX immediately searches the file for exact and for partial matches (using in the latter case a kind of `fuzzy logic'). If it finds an exact match, it then positions the cursor for the next field. If it finds only partial matches, it displays them for you to select one among them. If neither an exact nor a partial match is found, you are then asked to confirm that you want your input regarded as a `new' entry. In effect, INDEXX always operates in the `search' mode, looking for perfect and/or partial matches on the letters entered from the keyboard. For either type of match to be found, the leading letters must match. But what if you can't remember what the leading letters were? Suppose you remember having used the heading "apples", but cannot remember what the subheadings were. To retrieve all the subheadings for perusal and selection, you can use the wildcard, "*". You'd begin by retrieving "apples", e.g.: ap <Enter> (___) <Enter> (___) Having retrieved that heading, you can now use the wildcard, "*", to retrieve all its subheadings: * <Enter> (___) A list of all the subheadings (roughly sorted) will appear on the monitor. Use the cursor keys to select which one you want. We'll choose "varieties": - 16 - <Csr-down> (___) <Csr-down> (___) <Enter> (___) 148 <Enter> (___) We've just used the wildcard "*" to retrieve all the subheadings under the heading "apples". But the wildcard can be used `the other way around' as well, that is, to retrieve all the headings under which some particular subheading occurs. Suppose, for example, you wanted to add a further page citation, "301", to the subheading "sculpture", but could not remember under which heading - "art" or "museums" - you had previously listed this item. You could type "*" in place of the heading, enter "s" for the subheading, and leave the rest to INDEXX: the program will find and display (for your choosing) all the headings under which "s..." occurs as a subheading. Try it: * <Enter> (___) s <Enter> (___) <Csr-down> (___) <Enter> (___) 301 <Enter> (___) Notice that we had forgotten about "sauce" as a subheading under "apples". Since we specified only "s" as the subheading, INDEXX retrieved more partial matches than it would had we been more specific, e.g. if we had entered "scu" as the subheading. Obviously, the more specific the data entered, the fewer the number of partial matches. There is one restriction on the use of the wildcard: it may be used in place of EITHER the heading OR the subheading; but you cannot use it for both. If you think about it, using the wildcard twice would be equivalent to seeking a partial match on every record in the file. Instead, if you want to view the entire file, use function key <F2>. There are four more features of the Option 3 you'll want to test in this tutorial: typeface codes; `as-if' codes; cross- referencing; and on-screen editing. INDEXX utilizes the less-than and greater-than signs (accompanied by any letter) as typeface codes. For example, if you wanted The Fall of the House of Usher to print in italics, you could enter (in the string mode): i<The Fall of the House of Usher>i Note that codes come in pairs; a typeface - italics, underlining, boldface, etc. - is turned on by "n<" and is turned off by ">n". The actual letters you use is up to you: these are proxies (or - 17 - stand-ins or dummies) as far as INDEXX is concerned. You will need to do a `global search and replace' on these codes, using your word-processor, in the final file INDEXX produces. I personally use "i< ... >i" for italics; "b< ... >b" for boldface; and "u< ... >u" for underlining. After INDEXX has produced its final print file, I use my word-processor to replace these stand- in codes with the ones which my word-processor uses for italics, boldface and underlining respectively. Let's enter "i<The Fall of the House of Usher>i" as an entry in our index. Just to see what happens, we will deliberately `forget' to switch to the string mode. i<The Fall of the House of Usher>i <Enter> (___) Operating in the normal data input mode, INDEXX has reformatted this to read "Usher>i, i<The Fall of the House of" and has given us an error message that the typeface codes do not match. (Indeed they don't!) Erase the error message and switch to the string mode: <Space-Bar> (___) <F1> (___) <Enter> (___) Y (___) <Enter> (___) 38 <Enter> (___) Any item or page citation can be italicized, etc. But if you want to use typeface codes with page citations, be sure that no commas occur within the codes, otherwise, e.g. "i<70, 16, 51>i" would be sorted to yield "16, 51>i, i<70". If you try to enter such potentially problematic data, INDEXX will catch this mistake, and will require that you fix it. In this instance, you'd have to enter: "i<70>i, i<16>i, i<51>i". You may, if you like, embed (or `nest') one typeface code inside of another. It is perfectly permissible, for example, to enter "i<b<Russell>b, Bertrand>i". You will not be permitted, however, to use overlapping codes, e.g. anything of the sort "i<b<Russell>i, Bertrand>b". INDEXX will produce an error message if you try to enter such data. Sometimes it is necessary to inform INDEXX to sort (or alphabetize) some item `as-if' it were something else. E.g. you might want the movie title, "i<2001>i", alphabetized as if it were "two thousand one". INDEXX reserves curly brackets to designate `as-if' codes. In this instance, you'd type: <F1> (___) i<2001>i {two thousand one} <Enter> (___) Y (___) <Enter> (___) 159 <Enter> (___) - 18 - (I hope it's obvious why you needed to type <F1> to begin.) If, later, you wanted to add another page citation to this entry, you'd have to type as your heading "i<2>i". INDEXX would not find a match on "2" since the record in the file begins with "i", not "2". And you'd have to complete the pair of brackets, since INDEXX would reject, as a mistake, an entry of the form "i<...". Just to be sure you see how this is done, let's try it. Type: <F1> (___) i<2>i <Enter> (___) <Enter> (___) <Enter> (___) 219 <Enter> (___) (In this latter case, the use of the <F1> key wasn't strictly necessary; but it did no harm to use it.) One of the principal uses for `as-if' codes is to alphabetize correctly those pesky Roman numerals in Forewords and Prefaces. Suppose the page citations for "museums" includes both Roman and Arabic numerals. You could type the following: museums <Enter> (___) <Enter> (___) ix {0.09}, xviii {0.18}, 63 <Enter> (___) By using `as-if' values less than 1.00, you force the Roman numerals to appear earlier than any of the Arabic numerals (which, presumably, begin with "1"). If the largest Roman numeral is greater than nine, use two digits to the right of the decimal point. (You wouldn't want to use "{0.3}" and "{0.21}" as the `as-if' codes for "iii" and "xxi". If you did use these values, "iii" would be placed after "xxi" when they were sorted. Instead you'd want to use "{0.03}" for "iii" and "{0.21}" for "xxi" in order to get the order correct. To simplify matters, I myself make it a practice always to use two digits, "0.00" through "0.99", to the right of the decimal point when I use `as-if' codes for Roman numerals.) Incidentally, INDEXX will not permit you to `nest' typeface codes inside of `as-if' codes, nor `as-if' codes inside of typeface codes. INDEXX would report an error, for example, if you were to type: "i<2001 {two thousand one}>i". Instead, you would have to type this as was shown above, viz., "i<2001>i {two thousand one}" Sometimes you may want to cross-reference entries, e.g. you may want the entry "galleries" to direct the reader to the later entry "museums". To do this, you simply enter the cross- reference in the heading, and enter for the page number the special three-digit sequence "000". When INDEXX formats your - 19 - data file (in Option 5), if it finds that the only page citation is "000", it will not print any page citation whatever. Thus you would enter: <F1> (___) galleries. i<See>i museums <Enter> (___) Y (___) <Enter> (___) 000 <Enter> (___) If you want to use "see also", a case where you would want to include page numbers, you would do it this way: <F1> (___) magazines (i<see also>i journals) <Enter> (___) Y (___) <Enter> (___) 234, 267 (___) Sometimes one wants to place the "see also", not immediately after the heading, but at the tail end of the page citations. If so, you could enter the "See also" among the page citations and force it to appear at the tail end by using an `as-if' code, choosing for that code a number greater (e.g. 5000) than the last page number of the book. E.g. philosophy <Enter> (___) Y (___) <Enter> (___) 12, i<See also>i Hume$ David {5000}, 128 <Enter> (___) In this last entry you can see the use of a fairly common programming trick. The "$" in place of a comma between "Hume" and "David" is no mistake. INDEXX recognizes "$" as a `surrogate' comma if "$" occurs among the page citations. Among the page citations, the comma demarcates subfields, i.e. it separates one page citation from another. Thus the sort routine (Option 4) will reorder a string of page citations, "34, 12, 61" to read "12, 34, 61". But sometimes we want to use a comma, not as a marker between page citations, but as an integral part, i.e. within, a single page citation. In the current case, we want the final printout to feature a comma between "Hume" and "David", but we do not want the sort routine (Option 4) to treat "i<See also>i Hume, David {5000}" as being comprised of two separate page citations, viz. "i<See also>i Hume" and "David {5000}". Treated this latter (incorrect) way, the two parts would probably become detached from one another in sorting. So we use a "$" in place of the comma we want to `hide', in order to get the entire single string "i<See also>i Hume$ David {5000}" safely through the sort routine. Later, when it creates the final print file, INDEXX will replace all occurrences of "$" among page citations with commas. (Commas are not used as subfield demarcation in headings and subheadings, and there is consequently no need for the use of surrogate commas in either of those fields. I.e. "$" is regarded - 20 - as a surrogate comma only when it occurs among the page citations. [For more on cross-referencing within the page citations, see p. 40.]) There's one last feature to be discussed: the on-screen editing. To edit a record on-screen, you must first retrieve it from the file. To do this, you proceed exactly as if you were adding to it, except that instead of entering page numbers, you press <F4> to switch to the on-screen editor. For an example, let's suppose we decide, late in the day, that we want to go back and italicize the Roman numerals we earlier entered for "museums". We begin by retrieving that record, but instead of entering any page citations, we press <F4>: mu <Enter> (___) <Enter> (___) <Enter> (___) <F4> (___) The monitor is immediately switched to the On-screen Editing Menu (which also displays the current record being edited). The Editor Menu presents a number of editing options, including "Pages edit". Since it is the page field we wish to edit, we press "P": P (___) We can now use the cursor keys, etc., to move around in the field to edit it as we choose. Edit it so that it looks like this: i<ix>i {0.09}, i<xviii>i {0.18}, 21, 34, 55, 63 When you're finished, press <Enter> (___) to return you to the Editor Menu. What will fit of the revised field is displayed in the upper right hand corner. You're ready now to return to Option 3. Press "A" to accept the changes: A (___) Y (___) You are now back in Option 3. Use <F2> to view the file to see that the changes have actually been made: <F2> (___) V (___) <F10> (___) Once again you're back in Option 3, with the cursor positioned to accept page(s). Since we have no pages to enter (we simply wanted to correct an earlier entry), we'll erase the monitor display. Remember, <F6> will clear the monitor. - 21 - <F6> (___) (The on-screen editing facility can do a great many other tricks as well. But we'll leave that for you to read in the next chapter.) We've now seen most of what can be done in Option 3. Press function key <F10> to exit from Option 3: <F10> (___) Your data file will be written to disk, and with a press of the <Space-Bar> (___), you've climbed back up to the Main Menu. 2.7 Alphabetizing your index file Choose Option 4 - Alphabetize index (___). Once again the monitor clears and now you're in the Sorting option. All the settings are explained and illustrated on the monitor. You will probably want to look these over a bit, but we'll use the factory settings in this tutorial. When you're ready, press "Y" (___). A new screen will appear, listing all the `little' words to be ignored in sorting. Again, we'll use the factory settings. Press "Y" (___). The monitor will clear still again, and the sorting procedure will proceed. You'll be informed if there were any problem cases. There shouldn't be if you've followed the directions above correctly. Press the <Space-Bar> (___) to return to the Main Menu. 2.8 Formatting and printing your index file Now choose Option 5 - Format index for printing (___). Just as in Option 4, the settings will be displayed. Once again, we'll choose the factory defaults by pressing "Y" (___). Option 5 will now write out to disk a new file which contains the formatted output. You will be asked if you want to view the formatted file or if you want to print it. If you choose either one, you'll be given an opportunity - after it is completed - to choose the other as well. Press "V" (___) to view the formatted file. The monitor will clear, and the following should appear: - 22 - INDEX apples, 7, 45, 56, 77 pies, 75 sauce, 23 varieties, 85, 148 art sculpture, 85-93, 301 i<The Fall of the House of Usher>i, 38 galleries. i<See>i museums Hume, David, 49, 67 magazines (i<see also>i journals), 234, 267 museums, i<ix>i, i<xviii>i, 21, 34, 55, 63 philosophy, 12, 128. i<See also>i Hume, David Sullivan, Arthur, and William Gilbert, 23, 67 n.2, 100 i<2001>i, 159, 219 Windsor, Charles Philip Arthur George education, 75, 235 There are several things to note about this printout. - Some headings, e.g. "apples", have page citations, and others, e.g. "art", do not. If you enter page(s) for a heading, they will appear alongside that heading; if, however, you enter page(s) only for subheadings under a heading, the heading (as you can see in the case of "art") will appear without page citations. - The heading, "The Fall of the House of Usher", is alphabetized as if it began, not with "T", but with "F". This is so because "the" is among the `little words' the sort routine (Option 4) ignores. Indeed, this particular title was sorted as if it read "Fall House Usher" since "of", too, is among the `little words' ignored. - The movie title, "2001", was alphabetized as if it were "two thousand one". In this instance, we had explicitly specified, by using an `as-if' code in curly brackets, that it was to be so regarded. - All the explicit `as-if' codes ("{...}") and the surrogate commas ("$") have disappeared from this final printout. They were used in generating this output, but they do not occur in it. - In the entry "philosophy", the format routine has found "See also" among the page citations. The formatter is clever enough to figure out that when "See also" occurs among the page citations, it should be preceded by a period. Thus the formatter has changed the comma after "128" to a period. (Gosh, how clever can indexing software get?) - 23 - When you're through admiring your handiwork, press <F10> (___) to return you to Option 5. Again you're presented with the choice of viewing the file or printing it. This time choose the print option. Press "P" (___). Ready your printer, and then press the <Space-Bar> (___). The formatted output which you viewed a moment ago should now appear on your printer. When the printing is finished, press <F10> (___) to return you to the Main Menu. At the Main Menu, press 7 to exit (___). You will be warned that you've not made a backup copy. Normally you should make a backup copy, but since this is only a test run with faked data, you may safely ignore the warning. Press "N" [for "No"] (___). 2.9 Exploring the files on your data disk When you're back at the DOS level do a DIR[ectory] listing of your data disk. There should be three files on it: TEST-ONE.IND (this is your latest data file); TEST-ONE.OLD (this is the next- to-last version of your data file, it will contain your data as it appeared before it was input to Option 4); and third, TEST-ONE.PRT (this is the formatted output file, the one which you viewed and printed in Option 5). Finally let's return to INDEXX. At the DOS prompt, type: INDEXX <Enter>. 2.10 Retrieving a file in progress Again you're at the Main Menu. But unlike the first time you loaded INDEXX, there is now a data file on your data disk. Let's retrieve that file and continue adding data to it. Choose Option 2. Again you will be asked the drive and path for the subdirectory where your data file(s) are to be found. Specify the appropriate path (e.g. "A:\", "C:\BOOK\CHAPTERS\", etc.). The names of all your latest data files (those bearing extensions ".IND" or ".2ND") will appear on the monitor. (Note that TEST-ONE.OLD [your next-to-last file] and TEST-ONE.PRT [the formatted print file] are not included.) Since there is only one .IND file on your disk (and no .2ND), it alone appears. If there were several .IND files, all their names would appear, and you could choose one by moving the monitor highlighting with the cursor keys. But for the moment, you can only hit <Enter>. INDEXX will retrieve the file whose name is highlighted. You'll be asked if you want the data checked as it is read-in. If you've not modified the file with your word-processor, choose "N". If, however, you have modified the file, you'd be much safer choosing "Y", to see whether you may have, inadvertently, introduced some critical error. Once the file has been successfully read, the monitor will clear and you will be - 24 - informed of the status, etc., of that file. Pressing the <Space-Bar> will return you to the Main Menu. Choose Option 3. You can now continue adding data to your file as if you had never exited Option 3 earlier. As soon as you do however, note that the status changes from SORTED to UNSORTED. * * * * * From here on, you're on your own. Experiment a bit - especially try entering incomplete data (as above) to see how INDEXX returns partial matches; enter invalid data, e.g. use incomplete or overlapping typeface codes; use <F2> to view your file as you add data to it; view your file (using <F2> in Option 3) before and immediately after sorting it in Option 4; choose Word-by-word sorting in Option 4; etc. - and then when you begin to feel comfortable with INDEXX, read the following chapter which will discuss the operation of INDEXX in much greater detail. - 25 - Chapter 3 Detailed operation of INDEXX You will find this chapter easier to understand if you have worked your way through the tutorial in the previous chapter. 3.1 The Main Menu On completing any of Options 0 through 6 which you call from the Main Menu, you will automatically be returned to the Main Menu. You can, however, at several opportunities, stop any Option in mid-course, and return to the Main Menu, by pressing <F10>. There is a reminder to this effect on the bottom line of the monitor whenever <F10> is enabled. <F10> is always inoperative (disabled) during all reading and writing to disk. This is to insure that files are completely read and written. 3.2 Option 0 - Changing the defaults Option 0 may be selected from the Main Menu as well as from Option 3. If your computer is equipped with a color adapter card (any of CGA, EGA, or VGA)[4], and if you had at the outset chosen "F" (Full-color monitor) (see p. 8), Option 0 will present you with an intermediate menu offering (at first) two choices: (C)hoose Screen Colors; (S)et Operating Parameters. If you are using a monochrome monitor, this intermediate menu will be skipped and you will go directly to the `Set Operating Parameters' menu. ----------- 4. If your computer is equipped with both a color (graphics) card and a monochrome (text) card, before loading INDEXX you will have to choose one or the other by use of the DOS MODE command. At the DOS prompt, type MODE CO80 or MODE MONO to choose the Color or Monochrome adapter card, respectively. - 26 - 3.2.1 Choose screen colors (full-color monitor only) If you are using a full-color monitor and select (C)hoose Screen Colors, you will then be given a choice of (F)actory settings or (P)alette. If you choose Palette, you will then be presented with a display of 240 available screen colors and will be asked to choose a `primary' pair and a `secondary' pair. Just select numbers from among those displayed. Each pair consists of a foreground color and a background color. Foreground must not equal background otherwise what is printed on the screen would be invisible. The primary pair will be used for normal screen display; the secondary for highlighting (contrast). The primary pair must differ from the secondary pair. 3.2.2 Set operating parameters The `Set Operating Parameters' menu will list five parameters (six, if you are using a monochrome monitor) along with their current values which you may change if you like. To leave a value unchanged, press <Enter>, and the cursor will skip to the next parameter in the list. To change a parameter, simply type the new value. The five (/six) parameters which may be changed are: SOUND At various places, if INDEXX detects an error in data that is input from the keyboard, or an error in reading or writing to disk, it will emit a warning beep. If you wish to turn the sound off just press N [for No]; press Y[es] for sound on. FREQUENCY OF WRITING TO DISK In Option 3, INDEXX saves your data to disk (using the extension ".IND") after every tenth entry (or at whatever frequency [10 - 50] you specify in this Option). This automatic writing to disk assures that in the event of a power failure, or removing the disk prematurely, you will still have most of your work saved. On retrieving your file (Option 2), you will automatically be informed (on the start-up of Option 3) of the last entry recorded on disk, and thus should be able to pick up from that point with no difficulty. In Option 3 (as in Option 4 [alphabetizing of your data]), whenever INDEXX writes your file with the extension ".IND" to disk, it saves your previous version of the file by changing its extension to ".OLD". In this way, your next-to-last file is also always available. Should anything untoward happen to your ".IND" file, or to its backup, ".2ND", you can then use DOS to change the extension of the ".OLD" file to ".IND" and - at worst - you will be only ten (or so) items back from where you - 27 - last left off entering data. Normally, however, you will never have to concern yourself with the existence of the file ".OLD". It is created purely as an extraordinary backup measure. DATA ENTRY MODE (in Option 3) In Option 3, there are - as you saw in the Tutorial - two data input modes: Normal and String. Option 0 allows you to choose which will be the default data entry mode for headings. (The default mode for subheadings is always `string' [although it can be changed temporarily by <F1>]; and only the string mode is available for page entries.) If you choose `Normal' or `String' as the default mode, INDEXX will begin in that mode for each new data entry, i.e. for the entry of the heading, even if you had switched modes (using <F1>) on the previously entered heading. If, however, you choose `Manual', INDEXX will, each time, revert for the heading to the mode you used previously for the heading. In manual mode, you could thus switch back and forth between `Normal' and `String' by using <F1>, but would stay in either mode (for headings) until manually switching out of that mode. With a little experimentation, you will quickly find which setting best suits your data. (Remember, you can change this default `on the fly' by calling Option 0 from Option 3.) CONFIRMING UNMATCHED ITEMS (in Option 3) As you enter data in Option 3, some of it will be matched to previously entered data, but some of that data will be wholly new, i.e. not even partially matched to previous data. You have a choice: either new items will be added directly to the file, or INDEXX will pause and ask you whether you want the new item added. The pause allows you to double-check the spelling and correctness of the item. SCREEN UPDATE [Color adapters (CGA, EGA, or VGA) only] INDEXX can write data to the screen at two different speeds. On some vintage CGAs, writing at fast speed to the screen causes visible and objectionable `snow'. If you have CG adapter (CGA), experiment with the Fast speed to see what happens. If you have an EG adapter (EGA) or a VG adapter (VGA), choose Fast. SCREEN FOREGROUND [Monochrome monitors only] You can choose to have the monitor set to W(hite) foreground on a black background, or B(lack) foreground on a white background. (Note that changing this parameter does not take effect until the <Enter> key is pressed to exit the Default menu.) - 28 - Note that if you have a monochrome monitor attached to a color adapter (CGA, EGA, or VGA), you can choose either of Fast/Slow AND either of White/Black. On exiting Option 0, you will be returned to wherever you had been in INDEXX, viz. either to the Main Menu or to Option 3. 3.2.3 Command line parameters Any of the parameters which can be (re-)set in the `Set Operating Parameters' menu, may equally well be set at the DOS level, when loading INDEXX. In addition, you can set the path to be used for input and output. At the DOS prompt, one merely appends the relevant instruction to the command "INDEXX". For example, to specify that the path for input/output to be used for Options 1 and 2 is to be "C:\BOOK\CHAPTERS\", and the frequency of saving is to be 25, one could instruct DOS as follows: > INDEXX PATH C:\BOOK\CHAPTERS\, FREQ 25 The effect is exactly the same as having set the path in Options 1 or 2 and having set the FREQ(uency of saving) in Option 0. The permitted Command Line parameters are: Any of: PATH A:\...\...\ PATH B:\...\...\ etc. Either of: SOUND YES [or SOUND ON] SOUND NO [or SOUND OFF] Any value between: FREQ 10 - FREQ 50 Any one of: NORMAL STRING MANUAL Either of: CONFIRM YES [or CONFIRM ON] CONFIRM NO [or CONFIRM OFF] Any one of: WHITE BLACK COLOR [or RGB] Either of: FAST [applies to color adapters SLOW (CGA, EGA, VGA) only] If you are using a color adapter and specify any of "White", "Black" or "Color" on the Command Line, the initial screen which asks which kind of monitor is being used will be skipped. - 29 - The Command Line may be typed in any combination of upper and lowercase that you like. You may also use any (combination) of four different separators (delimiters) - spaces, commas, slashes, and/or dashes - between the various parameters. (Note "PATH", "SOUND", FREQ", and "CONFIRM" must immediately be followed by a blank.) And the various parameters may appear in any order. 3.2.4 Using a .BAT file If you use a standard set of parameters, you can set up a batch file which automatically specifies these parameters to INDEXX. You could use EDLIN or your word-processor to create a batch file, let's say, GO.BAT, whose content, for example, might be: INDEXX FREQ 25 -Path C:\book\Chapters\, black /fast Were you then to instruct DOS >GO <Enter> DOS would load INDEXX, would switch the monochrome monitor to display black on white, would write to the monitor adapter at the fast speed, and would set the two operating parameters, FREQ and PATH, to 25 and C:\BOOK\CHAPTERS\, respectively. If you don't specify any Command Line parameters, the defaults (start-up settings) are: Sound on FREQ 10 Normal Confirm on Slow White There is no default setting for PATH. 3.3 Option 1 - Creating an index file You will first be asked for the name of a path to a pre-existing subdirectory. (INDEXX cannot create a new subdirectory.) Once you have entered a path (e.g. C:\BOOK\CHAPTERS\), INDEXX will check to see that the designated subdirectory actually exists on the designated drive. If that path does not exist, an error message will appear and you will have to enter a new path. Once the path has been verified, you will next be asked to enter a name to be given to your file. The file name may contain up to eight letters. Blanks will be replaced by underscores, "_", to accord with IBM DOS conventions for naming. You can have several indexes in progress at one time on one disk. You might find it convenient, for example, to create one file per chapter: this would allow you to review and check small - 30 - segments as you work. These segments can later be assembled into a complete index (the procedure is described in Chapter 4). You might want to use names such as "CHAP01", "CHAP02", etc. INDEXX will provide appropriate extensions as needed for the various working files it generates. For example, when you first create a file, e.g. "CHAP06", INDEXX will add the extension ".IND" [INDex], yielding "CHAP06.IND". Since INDEXX holds only one file in memory at a time, creating a new file (with Option 1) or retrieving a file (Option 2) will clear from computer memory any file you had just been working on. To prevent your losing data, INDEXX is designed to prevent your getting to Options 1 or 2 without your updated file in memory having been written out to disk. 3.4 Option 2 - Retrieving a file in progress Just as in Option 1, you will have to enter the path to the subdirectory holding your .IND files. Option 2 will present you with a list of all files in the designated subdirectory which bear extensions, ".IND", and ".2ND". Any of these may be retrieved for updating (Option 3) or sorting (Option 4). Only sorted files (those having been alphabetized by Option 4) may be formatted for printing (Option 5). Choose the file you wish to retrieve by using the cursor keys. Then press <Enter>. You will then be asked if you want the records checked for errors as they are read. If you have altered the file in any way with your word-processor, you should choose "Y"; the reading will be slower, but critical errors will be detected and revealed. Once the entire file has been read into memory, its status - Empty, Sorted, or Unsorted - will be displayed (see section 3.6, page 48 for definitions). You will also be told the number of records currently in the file, and the day and time it was last modified. 3.4.1 Errors detected in reading the data file If Option 2 cannot read your file properly, it will display an error message identifying the particular record which contains the error and will tell you exactly what that error is. This error will have to be repaired with your word-processor. Similarly if you have merged two files (using Option 8) so that the resulting file contains more than 2,000 records or exceeds the memory capacity of INDEXX, Option 2 will cease reading the file when either of these limits is reached. You will then have - 31 - to use Option 10 in INDXPLUS to split the oversized file into smaller, more manageable, files. 3.5 Option 3 - Updating the index This is the workhorse section of INDEXX, where you will enter and edit your data. Two input modes are provided, for `normal' entries, and for `string' entries. 3.5.1 Normal data entry mode On first entering Option 3, the monitor will be set to receive data in the so-called `Normal Mode'. You can switch between modes by pressing <F1>. The normal mode is ideally suited for all data which are either (1) a single word, e.g. "tribalism", or (2) a person's name, e.g. "Henry James" or "C.E.M. Joad". All of the following are acceptable entries in the normal mode: C. E. M. Joad C.E.M. Joad Henry James tribalism In the `normal' mode, INDEXX looks for the last blank space and considers the single word following it to be the surname to be alphabetized; what precedes the blank is considered the first, or given, name(s). The series of words will be re-formatted accordingly, e.g. "Henry James" will become "James, Henry". Given names may, of course, include, or consist entirely, of initials. Although perfectly suited for single words and the names of single authors, etc., the normal mode is ill-suited for entries which consist of two or more words, e.g. "New Jersey" or for a series of names of two or more persons, e.g. "William Gilbert and Arthur Sullivan". For the normal mode would re-format the former of these (incorrectly) as "Jersey, New" and the latter (again incorrectly) as "Sullivan, William Gilbert and Arthur". To accommodate such cases, you will need to take recourse to the `string' mode. 3.5.2 String data entry mode To enter data which do not fit the `normal' pattern, i.e. for which no re-ordering of words is wanted, press <F1> to switch to - 32 - `string' input.[5] In the string entry mode, you would enter e.g. "New Jersey" or "Gilbert, William, and Arthur Sullivan". Thus if you want to enter a person's name in the string entry mode, and want INDEXX to alphabetize on that person's surname, it will be necessary - unlike the `normal' entry mode - for you, yourself, to place the surname before the person's first name. In the normal entry mode, you would enter "Henry James"; but in the string mode, you would want to enter, instead, "James, Henry" [with a comma]. 3.5.3 Subheadings Once you have entered the heading, the cursor will jump down on the screen ready for you to enter a subheading. If there is to be no subheading, simply press <Enter> to skip. For example, after you have entered "James, Henry" as a heading, the cursor will jump down ready for you to enter a subheading. You might then choose to enter: "education". The default entry mode for subheadings is permanently set to String. You can, however, switch to Normal with the use of the <F1> function-key. 3.5.4 Internal file fields If you were to enter the data shown above (along with some suitable page citations), the list as internally stored in computer memory - and on writing (extension ".IND") to disk - might look like this: "Joad, C.E.M.","","64, 82" "James, Henry","education","42" "tribalism","","81, 88-91, 46 n2" There are a few features worth pointing out in this last list. - You will see that a data record in INDEXX always consists of three fields - a heading, a(n optional) subheading, and a list of page(s) - each field defined by a pair of double quotation marks. Because double quotation marks are used in this way, as markers (delimiters), defining the beginning and end of each field, they cannot be allowed within a field. Thus if you use double quotation marks in your data, these will be converted internally by INDEXX to single quotation marks, i.e. " ` " and " ' ". ----------- 5. You can make the string mode the default mode for headings by using Option 0. See p. 28. - 33 - - Note that INDEXX has closed the spaces in Joad's initials: "C. E. M." has been re-written as "C.E.M."[6] - In this example (prior to Option 4), the data have not been alphabetized, nor have the page entries been sorted. The data have been accumulated simply in the order in which they were originally entered. Matches have been found, but that is all. 3.5.5 Partial matches In the file fragment given immediately above, we have a heading reading "James, Henry". Suppose now you were to enter "H. James" in the normal mode (or, equivalently, "James, H." in the string mode). INDEXX would search the file and would find a match on "James". But the given names ("H." and "Henry", respectively) are different. Is James, H. the same person as James, Henry? INDEXX will regard this as a `partial match' and will ask you to decide. You will be presented with one to five `pages' listing all the partial matches (to a maximum of fifty). You can select any one of the displayed choices by moving the on-screen highlighting to it and pressing <Enter>. If you were to select the match on "James, Henry", the entry would be recorded using the longer of the given names (in this case "Henry"). Only ten partial matches are displayed per `page'. If there are more than ten partial matches, you can switch pages by pressing <PgUp> for the previous page, or <PgDn> for the next page; or by pressing the <Up-arrow> key (for the previous page) when the highlighting is on the topmost item, or the <Down-arrow> key (for the next page) when the highlighting is on the bottommost item. If you reject all partial matches by pressing the <F10> key, you will be asked whether the unmatched datum is to be entered as a new record in the file. If you answer "N(o)", you will be returned to the original screen in Option 3 with your input data intact. You can then erase it by the use of the <Esc>-key, or you can modify it. (See subsection "Editing the input line", p. 42 below.) To be explicit: According to the fuzzy logic adopted in INDEXX, Two items will `partially match' if either one is the so-called ----------- 6. This compacting of initials saves computer storage and provides a single, uniform format for comparing names. But not all publishers favor this format. Option 5 provides the opportunity at the time of formatting the index for printing to (re-)introduce spaces between initials. - 34 - proper first part of the other. Thus "J" and "James" partially match; so do "individuals" and "ind". Suppose, for the sake of illustration, there are two Boschs in your index: Herbert and Hieronymus. To add a page citation to the entry for the former, you might enter "He B" in the normal mode (or "B, He" in the string mode) and INDEXX would find a partial match only for "Bosch, Herbert". (That is, "He" partially matches "Herbert", but does not match at all "Hieronymus".) If, however, you were to enter "H B" in the normal mode (or, equivalently, "B, H" in the string mode), INDEXX would find two matches and you'd have to choose between them. Let's take another example. Suppose your file were to contain both "idea" and "idealism" as distinct headings, and you were to enter "idea". INDEXX would find a partial match on "idealism" but a perfect match on "idea", and thus would not present "idealism" as a possible match for your choice. INDEXX would instead report a perfect match on "idea". Thus if your file contains both "idea" and "idealism", to retrieve the latter you could enter any of: "i", "id", "ide", "ideal", "ideali", "idealis" or "idealism"; but you could not retrieve "idealism" by entering "idea". Any time you enter an item which partially matches an earlier one, but which is longer than that earlier one, you'll be asked if this is correct, and if you confirm that it is correct, the earlier one will be updated to the new, longer version. Thus suppose the file already contains "Bosch, H.", and you were now to enter "Hieronymus Bos" in the normal mode (or "Bos, Hieronymus" in the string mode.) You would first be given the opportunity to select the earlier "Bosch, H." as your intended entry. If you select that particular match from the partial matches presented, you will then be asked to confirm that you want the earlier item revised to this new, longer version. Should you now reply "Y", INDEXX would revise the entry in the file. The resulting heading would then read "Bosch, Hieronymus", i.e. would be the result of having combined parts of the file data and parts of your newly entered keyboard data. It is important to note that searches for matches and partial matches are `case sensitive', i.e. sensitive to the difference between upper and lowercase. Although you may take shortcuts with names - e.g. entering "Bo" or "Hi Bo" for "Hieronymus Bosch") [in the normal mode] - such names must consistently begin with a capital or lowercase letter. Thus "E.e. cummings" (entered in the normal mode) will not match "cummings, e.e." [and "e.e. cummings" will not match "Cummings, e.e."]. However, any of "cu", "e cummi", "e. cummings", "e. e c", and "e.e. cummings", will return a partial match with "cummings, e.e.". - 35 - 3.5.6 Wildcard searches Suppose you want to review all the subheadings under some particular heading, or suppose you've forgotten under which heading(s) you earlier placed the item "interpretation", you can use the so-called "wildcard" - the asterisk "*" - to retrieve the data. If you wish to review all the subheadings under some particular heading, retrieve that heading in the standard manner, e.g. by typing a few of the leading letters and allowing INDEXX to find a partial match. Then, for the subheading, type the wildcard, "*". All the subheadings (to a maximum of fifty) under that heading will be displayed in alphabetical order. For example, to review all the subheadings under the heading "case studies", you might begin by typing "case stu" for the heading, and then when asked for the subheading, you could type "*". Or suppose you want to search `the other way around', i.e. list all the headings under which some particular subheading occurs. You could enter the wildcard in place of the heading, and enter the particular subheading at the second stage. For example, the subheading "history of" might occur under several different headings in an index. To list all the headings under which "history of" occurs as a subheading, you could enter "*" for the heading, and "histo" for the subheading. In short, the wildcard can be used in place of either the heading or the subheading. Naturally, you cannot use the asterisk twice in one search, for that would be equivalent to requesting matches on every record in the file. Instead, if you wish to review the entire file, use function key <F2>. 3.5.7 Typefaces: italics, boldface, etc. INDEXX reserves the greater-than and the less-than signs, in combination with a letter, to signify typeface codes. E.g. to print "a priori" in italics, you might enter: "i<a priori>i". Typeface codes must occur in matched pairs; you may use any letters you like. These codes are simply proxies, i.e. they `reserve a place' for the actual codes recognized by your word- processor. They have to be replaced in the final print file by your word-processor. I, personally, use "i<...>i" for italics, "b<...>b" for boldface, and "u<...>u" for underlining. Typeface codes may be nested inside of one another; but they must not overlap. Typeface codes cannot occur within, or contain, `as-if' codes (see next subsection). For more examples, see the Tutorial, p. 18. - 36 - 3.5.8 `As-if' codes (forced sorting) Occasionally you will want to direct the sorting routine (Option 4) to alphabetize a word, or sort a number, `as-if' it were something else, e.g. to alphabetize "i<2001>i" as if it were "Two thousand one". To force Option 4 to alphabetize differently, simply place the desired alternative in curly brackets immediately after the words (or numbers) to be treated specially, e.g. "i<2001>i {two thousand one}". INDEXX will then sort this `as-if' it were "two thousand one" rather than as "2001". This technique is particularly useful for sorting Roman numerals (e.g. "i<iv>i {0.04}") and for forcing certain cross- references (e.g. "i<See also>i ... ") to the tail end of a string of page citations. (For examples, see the Tutorial, pp. 18-21, and later in this chapter, p. 40.) Using `as-if' codes is particularly useful, too, if you are creating both a Names Index and a Subject Index. If you select a so-called `Sort 2' in Option 4, INDEXX will alphabetize your data in two separate lists: one containing all records beginning with an uppercase letter, and a second one containing those records beginning with a lowercase letter (roughly a Names Index and a Subject index). But occasionally a name might begin with a lowercase letter (e.g. "cummings, e.e.") and a subject with an uppercase letter (e.g. "Thomism" or "SETI"). In such instances, you can direct INDEXX to sort these records `as if' they were something else. E.g. cummings, e.e. {Cummings e.e.} Thomism {thomism} SETI {seti} `2001' {two thousand one} Note: Commas are not permitted within `as-if' codes. 3.5.9 Foreign alphabets On IBM computers and close-compatibles, foreign language letters may be entered from the keyboard by holding down the <Alt> key and typing the letter's three-digit ASCII code on the numeric keypad. For the foreign letter ASCII codes, consult the User's Manual for your computer. Option 4 will later ask you whether you want character 225 alphabetized as a Greek Beta or as a German Eszett. - 37 - 3.5.10 Cross-referencing Cross-references may occur either within the heading or subheading fields, or among (typically at the tail-end of) the page citations. Examples: galleries. i<See>i museums magazines (i<see also>i journals), 25, 67 n.2, 207 Gaposchkin, Cecilia, 45, 89. i<See also>i Payne, Cecilia In the first case, there are no page citations. You would enter (in the string mode) the line shown as a heading (or subheading as the case may be) and then enter the dummy, special, page code "000". When Option 5 encounters the three-digit code "000" occurring all by itself, it suppresses printing of the page citations. The second example is the most straightforward. You simply type the heading (or subheading) exactly as illustrated (in the string mode) and enter the page citations in the page field. In the third example, the cross-reference occurs among the page citations. This kind of case will be examined separately below (on p. 40). 3.5.11 Page citations Remember (from the Tutorial, p. 19) that the three-digit sequence "000" is reserved as a special code to be used in place of bona fide page citations. You use it to signal the formatting routine (Option 5) that the heading or subheading consists entirely of a cross-reference of the kind "xxx. i<See>i zzz" and that no page numbers whatever are to be attached. Page citations often contain more than just page numbers. For example, along with ordinary page numbers ("12, 26"), one may find "Chapter 4", "101 Table 2", "Appendix II", "Section 9.3.8", "i<See also>i natural kinds", etc. The question arises how to sort subfields (see pp. 12 and 20) containing such `mixed' data: these subfields cannot simply be sorted by numerical value (many citations are not numerical, e.g. "Preface") nor can they simply be sorted alphabetically (for then we would have this order: 1059, 14, 162, 21, 298, 47, 500, 63). The technique used in INDEXX is to sort first by the value of the leading number if the citation (i.e. the subfield) begins with a number (e.g. "252" or "101 Table II"), otherwise to sort alphabetically (treating upper- and lowercase as equivalent) in accord with the so-called ASCII equivalent of each character of - 38 - the subfield (e.g. "i<See also>i Vulcanism").[7] The ASCII equivalents are built into the computer. The partial listing is: [space] 32 ! 33 " 34 disallowed within fields in .IND files # 35 used as error flag on page fields in .IND files $ 36 used as a `surrogate comma' within page fields in INDEXX % & ' ( ) 37-41 * 42 reserved for wildcard searches in INDEXX + , - . / 43-47 0-9 48-57 : ; 58-59 < 60 reserved as a typeface code in INDEXX = 61 > 62 reserved as a typeface code in INDEXX ? @ 63-64 A-Z 65-90 [ \ ] ^ _ ` 91-96 a-z 97-122 treated as 65-90 when occurring among page citations in an INDEXX file { 123 reserved as an `as-if' code in INDEXX | 124 } 125 reserved as an `as-if' code in INDEXX ~ 126 Thus, for example, the page field Preface, 25, 25 n.3, Chapter 6, Appendix II would be sorted this way: 25, 25 n.3, Appendix II, Chapter 6, Preface ----------- 7. There is one exception. Suppose both a range of pages and a footnote citation in the same field begin with the same page number, e.g. suppose the page citations in a given record were to contain both "233-6" and "233 n.4". An ASCII sort would place the latter before the former since the space in the latter has a lower ASCII value (32) than the dash (45) in the former. But INDEXX has been programmed to reverse these two values in sorting page citations, and thus would place the span of pages before the footnote reference. - 39 - You may well prefer, however, that these subfields be sorted, not in this quasi-alphabetical order, but in some other order, e.g. reflecting the sequence in which they occur in the book, i.e. you might want "Preface" occurring first, and "Appendix II" appearing last. The technique for `forced' sorting is quite simple. You just append an `as-if' code (see above, p. 37) to the subfield, and that subfield will be sorted `as-if' it were the code specified. To force "Preface" to the front of the list, and "Appendix II" to the end, you could enter this data (or edit it on-screen) so that it reads: Preface {0.05}, 25, 25 n.3, Chapter 6 {234}, Appendix II {5000} 3.5.11.1 Cross-references among page citations There is one caution, however. Beware of using commas within a given subfield. Remember: commas are used to separate the various subfields within the page field of a record. Take care, then, not to enter a cross-reference of this type among the page citations: i<See also>i Logic, history of; Mathematics, sources of {5000} (The `as-if' code "{5000}" has been used here to force this cross-reference to the tail-end of the page citations.) The sorting routine would interpret this string as three distinct subfields - "i<See also>i Logic", "history of; Mathematics", and "sources of {5000}" - and would sort it as: history of; Mathematics, i<See also>i Logic, sources of {5000} Instead, you should use a `surrogate' comma (i.e. "$") which the formatting routine (Option 5) will recognize as a comma, but the sorting routine (Option 4) will not. Using surrogate commas within the subfield we wish to `hold together', the complete page field on some given record might then look like this: 25, i<See also>i Logic$ history of; Mathematics$ sources of {5000}, 60, xiv {0.14} The sorting routine (Option 4) would then re-order these subfields this way: xiv {0.14}, 25, 60, i<See also>i Logic$ history of; Mathematics$ sources of {5000} - 40 - And the formatting routine (Option 5), which suppresses `as-if' codes and replaces surrogate commas with real ones, would then print the field as: xiv, 25, 60. i<See also>i Logic, history of; Mathematics, sources of If you look closely, you will see that Option 5 actually does one other neat trick as well. Option 5 `looks at' the punctuation mark immediately before "See also" among the page citations and changes it to a period (in this instance changing the comma after "60" to a period). (This special feature works only if "See also" occurs among the page citations; it does not work, nor should we want it to, if "see also" occurs in a heading or a subheading, i.e. before the list of page citations. See examples on p. 38.) 3.5.12 Automode If you are working your way systematically, page by page, through a book, you may find yourself entering several references on a specific page. In such a case you might like the page numbers to be entered automatically, to be changed only when you, in fact, move on to another page. Such a means is available with the Automode. To switch the Automode on or off, use <F3>. When the Automode is on, you can - but need not - supply a page number. If you do, the Automode will automatically update to the page number you enter; if you do not, the Automode will provide the page number, i.e. it will default to the last page you entered. To change the default page you can simply enter data alongside the "[Page(s)]:" prompt; or you can increment the Automode by one page by pressing <F5>; or decrement it by one page by pressing <F7>. If you hold down <F5> or <F7>, the Automode counter will advance at the speed which the keys repeat: after an initial pause of one-half second, at a speed of 10 characters per second.[8] The use of the three keys - <F3>, <F5> and <F7> - is shown on the monitor during the operation of Option 3. The use of <F3> is always shown; the use of <F5> and <F7> only when the Automode is switched on. The status of the Automode (whether on or off) is shown constantly on line seven, but when the Automode is switched on, the page prompt carries a second reminder in that it will appear in square brackets (e.g. will appear as "[Page(s)]:"). The square brackets indicate that the page entries are optional. ----------- 8. This feature is hardware-dependent, and is not available on some clones. - 41 - The Automode will switch off if you enter alphabetic data, or a negative or a decimal number for a page citation, e.g. would switch off if you entered "i<xiv>i {0.14}". If <F3> does not turn on the Automode, check your last entry - which is displayed on line four - and it will show that you entered alphabetic data or a negative or a decimal number for the page citation. 3.5.13 Editing Two different editing facilities are available in INDEXX: (1) you can edit the data as you enter it from the keyboard, i.e. before it is moved into the file (database); and (2) you can retrieve any record from the file and edit it on-screen. Any field - heading, subheading, pages - entered on-screen in Option 3 may be up to 255 characters in length for a total record length of 765 characters. (This figure of 765 applies only to the length of fields input from the keyboard. Individual fields in records may exceed these figures by (1) your modifying them with a word-processor, (2) by your successively adding more data to the list of page citations (the 3rd field of a record), or (3) by your retrieving a record and editing its fields on-screen. The editor for this latter function can handle up to 1000 characters per field, i.e. 3000 characters per record.) The editing facilities in INDEXX make use of the following editing keys: <Esc> erase current field <Backspace> erase character to left of cursor <Del> erase character at cursor <Ins> toggle (switch between) the Insert and the Overwrite modes <Rt-arrow> move cursor right one character <Lt-arrow> move cursor left one character <Up-arrow> move cursor to same column in line above <Dn-arrow> move cursor to same column in next line <Ctrl><Rt-arrow> move right one word <Ctrl><Lt-arrow> move left one word <Home> jump to left edge of current line <End> jump to right edge of current line <PgUp> jump to beginning of field <PgDn> jump to end of field This list, defining the editing keys, is available within Option 3 in an on-screen `help window' simply by your pressing function key <F8>. To erase the help window, press function key <F10>. In the insert mode (designated by a full-height non-blinking cursor), the data typed on the keyboard are inserted into your text and previously typed data are pushed to the right; in the - 42 - overwrite mode (designated by a half-height blinking cursor), your new data overlay the old. Do not press <Enter> (the carriage-return) as you near the right-edge of the monitor as you would with a typewriter. The keyboard routine in INDEXX will automatically `wrap' your text to the next line, breaking at a space between words. When a line wraps, a half-diamond will automatically appear at the right edge of the line so you can gauge its length. These half-diamonds are removed when the record is added to the .IND file. 3.5.13.1 Editing the input line In Option 3, as you enter data from the keyboard, you can edit it using any of the editing keys just described. However, once you've pressed the <Enter> key to move along to the next field, e.g. to skip down from entering a heading to entering a subheading, there's no `climbing back up' to the previous field. If you spot a mistake in the data you've entered on the monitor in a previous field, there are two remedies: (1) you can continue, i.e. enter the mistaken data into the file, and then, later, retrieve it and edit it (see the immediate following subsection); or, (2) you can press <F6>, thus erasing all input data from the monitor, and begin again. 3.5.13.2 Retrieving and editing a file record INDEXX can retrieve any record in your file for on-screen editing. To retrieve a record from the file, proceed as if you were going to add another page citation to it. That is, enter the heading (or the leading letters thereof or a wildcard) in the standard manner, and then the subheading. When it comes time to enter the page(s), instead press <F4>. This will call up the on- screen Editing Menu. The retrieved record will be displayed in a window at the upper left of the monitor; and as you edit the record, the altered fields will be displayed in the window at the upper right. Immediately below these two windows there will be a list of editing choices. Certain of these choices will always appear: Heading edit Remove record / block <F10> Cancel and return to Update screen Three others occur only as appropriate: Subheading edit Pages edit Accept changes as shown above, at right - 43 - Three of these six choices - Heading edit, Subheading edit, and Pages edit - are self-explanatory. If you choose any of these, the monitor will clear, displaying the field you have just selected, and you may proceed to edit that field (to a limit of 1000 characters). Pressing <Enter> will return you to the Editing Menu, where you will again be presented with the just- listed set of choices. You can then, if you like, proceed to edit another field. A bit of terminology must be explained. The term "block" refers to a set of records all sharing the same heading. For example, the following three records comprise a `block': "logic","symbols","23, 67" "logic","","10, 276" "logic","teaching of","45, 55-57" (If the file is not sorted, the individual records comprising a block may be scattered throughout the file.) Suppose you were to retrieve and edit one record from among a block, let's say, the first of those just shown. In particular you change the subheading "symbols" to "notation". The changed record now reads, "logic","notation","23, 67" The particular change you have made is unproblematic, and if you accept this change - by choosing "A" (for "Accept changes" on the Editing Menu) - INDEXX would alter this one record in the file accordingly. But suppose instead of changing the subheading "symbols" to "notation", you were to have altered the heading so that you have "logic (formal)","symbols","23, 67" If you alter the heading and accept the change, INDEXX will ask you whether you want to make this change in just the one record you have edited or whether this is to be `global' change to be made to all headings in the block. You will have to choose "S" (for "single record") or "B" (for "block"). There is one tricky case. Suppose all the records in the block sharing the heading "music" have subheadings, e.g. "music","history","68, 157" "music","performance","12, 202" and suppose you've called up on the monitor, for on-screen editing, the record whose heading is "music". Clearly there is no such record in the file. But INDEXX pretends that there is. You can recognize that this is a `pretend' record because both the subheading field and the page field will be missing in the display in the upper left window. Thus if you make a change - 44 - (e.g. changing "music" to "musical affairs"), when it comes time to accept the change, the only choice you will be offered is for an entire block (since there is no actual record corresponding to this `pretend' record). This sounds more complicated than it is. You don't actually have to understand this technical explanation. INDEXX is programmed to recognize the difference between bona fide records and `pretend' ones, and will guide you through the appropriate choices by altering the questions it poses to you. Similarly if you choose to remove [delete] a record, INDEXX will ask you whether you want to remove just the single record appearing on the monitor or whether you want to remove all records in the block. Again you will have to choose "S" (single record) or "B" (block). (If the record displayed is a `pretend' record, the only removal permitted is, obviously, for the entire block.) It is inevitable that computer programs reflect something of the personality of their designers. INDEXX is no exception. I happen to be a terribly cautious person. It is just not my nature to throw away chunks of data. (Sometimes we change our minds and want to salvage what we earlier saw fit to discard.) And so, whenever you instruct INDEXX to remove a record or a block from your file, the record(s) are not allowed to evaporate. They are added to the `trash bin' file on your data disk. This trash bin file is created automatically. It will bear the same file name as your data file (i.e. as your .IND file) but will have the extension ".OUT". If, for example, you were to remove records from the file TEST-ONE.IND, those records would be moved into the file TEST-ONE.OUT. If you ever wanted to rescue some among those discarded records, you could do so with your word- processor. Once you have accepted some changes in the editing routine, INDEXX will pause and flash the message "Please Wait". "What is it doing?", you might ask. INDEXX is searching your file, looking for redundancy. For every time you edit a heading or a subheading, there is the possibility of your creating a record which matches another one in the file. If, for example, you were to alter the record "Canada","politics","69" to now read "Canada","government","69" there is the possibility that this latter record would match another one in the file. If INDEXX finds such a match, it will combine the two records. Whenever you edit a file on-screen, however slightly, its status will revert to "Unsorted". - 45 - 3.5.13.3 Searching for records with error flags [###] If you inadvertently enter duplicate page citations on a single record - e.g. "12, 25, 34, 25, 62" - Option 4, when it sorts the page citations, will recognize the error, will tag the offending page field with a leading flag [###], and will not sort the page subfields on that record. You can retrieve and correct these flagged records by using the search capability of the editing facility in Option 3. In Option 3, when the cursor is positioned to accept input for either the heading or the subheading, if you type <F4>, you will then be presented with two further options: "S" (search for error flags), and "H" (help with editing). If you choose "S", the editing facility will search your file for successive records bearing error flags in their page fields. You will then be able to repair these records on-screen, one by one, using all the standard facilities of the editing facility. (You cannot search for error flags if the cursor on the update screen of Option 3 is positioned in the page field.) 3.5.14 Review of function key usage <F1> Switches between normal and string entry modes (available for headings and subheadings only; i.e. not available for the page field) <F2> Presents the data file for viewing. <F3> Turns on/off Automode. <F4> Switches from the Update to the on-screen Editing facility <F5> If Automode is on, increments page number. <F6> Erases all fields displayed on the monitor. (The <Esc> key, in contrast, erases only the current field, i.e. the field in which the cursor is positioned.) <F7> If Automode is on, decrements page number. <F8> Opens a window displaying the functions of the various editing keys. <F9> Calls Option 0 which allows you to change various default settings. <F10> Reverts to previous procedure, e.g. cancels the display of partial matches and reverts to main display for Option 3; or terminates viewing the file (initiated by <F2>); or closes the Help Window opened by <F8>; or writes the file to disk and reverts to the Main Menu; etc. The operation of all these function keys is shown on the monitor in Option 3. - 46 - 3.5.15 Error messages in Option 3 As you enter or edit data in Option 3, INDEXX checks to see if the entries violate stipulated formats or operating constraints, e.g. whether you have used two wildcards, whether you have used a dangling comma, whether you have used overlapping typeface codes, whether there are more than 50 partial matches, etc. These are all minor errors and you will be given the opportunity to re- enter your data in the correct format. But in addition, there are two kinds of problems which cannot be overcome by your modifying the offending item. Because all editing, sorting, and formatting is performed on files held entirely in memory, these files are limited by the available memory. Option 3 is designed to check to see that neither memory nor various critical parameters in the program are exceeded. 3.5.15.1 Message 1: "Only 2,000 entries permitted" The maximum number of records permitted in an INDEXX file is 2,000. You can keep an eye on how many records there are in your file by reading the figure under the heading "Records" on the monitor. If you try to add records beyond the maximum value of 2,000, an error message "Only 2,000 entries permitted" will appear. If you get this message, you can continue to add page citations to records already in the file, but you cannot add any new records. To construct a file which exceeds this limit, you will have to use Option 8 in the auxiliary program, INDXPLUS (see Chapter 4). 3.5.15.2 Message 2: "Out of memory" Data files are held in a complicated fashion in computer memory in both a large fixed block (exactly 180,000 bytes [characters]) and in an `overflow' block. The overflow is of necessity relatively small, some 12,250 or so bytes. The amount of `free' (overflow) memory is posted on the screen in Option 3 under the label "Bytes free". Don't be alarmed if the amount of free memory jumps around (both up and down) as you enter data. As long as there are 1000 or more bytes available everything is fine. If, however, your data records are on average unusually lengthy and exhaust the overflow area, an error message "Out of memory" will inform you that there is no memory remaining for further additions to your file. Again, to construct a file exceeding this limit, you will have to use Option 8 in INDXPLUS (see Chapter 4). - 47 - 3.6 Option 4 - Alphabetize index 3.6.1 Sort-1 / Sort-2 SORT-1 produces 1 list: it alphabetizes the data without regard to upper and lowercase initial letters. SORT-2 produces 2 lists: it will place all records beginning with capital letters at the head of the index, followed by all records beginning with lowercase letters. SORT-2 corresponds (roughly) to a so-called `Names/Subject' index. Suppose you were to have the following data: "Aristotle; zoology; aggression; Zamfir, G." The results of each sort would look like this: SORT-1 SORT-2 aggression Aristotle Aristotle Zamfir, G. Zamfir, G. zoology aggression zoology If you have subheadings in your index, the subheadings will themselves always be sorted in the manner of SORT-1, i.e. intermixed without regard to upper or lowercase (see the `musical recording' example on page 2 above). 3.6.2 Letter-by-letter / Word-by-word sorts In a letter-by-letter sort, blankspaces between words are ignored. In a word-by-word sort, a blankspace is alphabetized as if it were a letter `earlier than' "a". Examples: Letter-by-letter Word-by-word game game gamekeeper game plan game plan game theory gamete gamekeeper game theory gamete 3.6.3 Float / Sink Some editors want names beginning with "Mac" and "Mc" (e.g. "McBride") to be `floated' to the top of the entries beginning with "M" (or "m"). You can instruct INDEXX to `float' or to `sink' such names. The differences are illustrated on-screen in Option 4. - 48 - 3.6.4 Character 225 You can set ASCII character 225 to be alphabetized either as the Greek beta or the German eszett. 3.6.5 File header The foregoing four choices - Sort-1/Sort-2, Letter-by- letter/Word-by-word, Float/Sink, Beta/Eszett - are automatically recorded as the ninth record of the header in your file. 3.6.6 Ignoring `little' words Once you have chosen/confirmed the sorting parameters (see preceding paragraph) you will next be presented with a display of the `little' words to be ignored in sorting. The factory defaults are: "&, a, an, and, in, of, on, the". You can simply elect to use these defaults, or you can modify the list on- screen. To modify the list, press "N", and then follow the on- screen directions. Note the editing options displayed in the info-line (line 25). You can erase the entire list with the <F6> function key; re-set to the factory defaults with <F4>; etc. Using the factory defaults, INDEXX would alphabetize, e.g., "i<The Mill on the Floss>i" as if it were "Mill Floss", i.e. would ignore "The", "on" and "the". Note that although only "the" (lowercase) occurs in the list of words to be ignored, its presence there will serve to instruct INDEXX to ignore all of: "The", "the", and "THE". Of course you could have accomplished the same thing with explicit `as-if' codes. You could have entered in your data file: i<The Mill on the Floss>i {Mill Floss} The difference between `ignored' words and `as-if' codes is that `ignored' words apply to every record in the file while `as-if' codes apply to the one record to which they are appended. `As-if' codes always take precedence over `ignored' words. For example, suppose that for most cases you want "a" to be ignored, but not in the case of "a priori reasoning". You could instruct INDEXX not to ignore "a" in this latter case by entering in your data file: "a priori reasoning {a priori reasoning}". Using this `as-if' code would force INDEXX to alphabetize this record among the entries under "a" and not among the entries under "p". When you are through editing the list, you will be asked if you want to save the list in the file "IGNORE.DEF". This is a special, important file if you want to create more than one .IND file for a particular job. Suppose, for example, that you will be creating different files for the many chapters of a book with - 49 - the intention to combine them later using Option 8 in INDXPLUS. If so, you'll want to be sure that these files are sorted using exactly the same set of ignored words; if different sets are specified, Option 8 will not be able to combine the files. To make it easy to get exactly the same sets of ignored words, Option 1 - the Option used to create a new file - will look in the designated directory for the file IGNORE.DEF and will use it, if found, instead of the factory defaults in creating the file header for your new file. Thus you can easily assure that the same list of ignored words is carried from file to file as you create different files for a single job. Whether you choose to save the list of words to be ignored in the file IGNORE.DEF or not, the list itself will be entered as the eleventh record in the file header of the .IND file so that you need specify the list only once per file, i.e. is always available whenever you load your working file into INDEXX using Option 2. 3.6.7 Combining of records After you have confirmed/edited the list of words to be ignored, INDEXX will proceed to alphabetize your file. If you have edited your file with a word-processor, it is possible to have inadvertently created a duplicate record. Once INDEXX has sorted your file, it will make a pass through all the records searching for redundancy. If any records match, they will be combined and their page citations collated. You will be informed on-screen if any duplicate records have been found in the file. 3.6.8 Sorting page citations; Problem cases After Option 4 alphabetizes your data, it will make a final pass, checking the page citations on each reference. If these are out of order, they will be re-ordered in the manner previously described on pp. 38-40. If in sorting the page citations on an entry, two identical citations are found, that entry will be flagged with three sharps [###], and you will be informed on the monitor of the number of such problem cases.[9] ----------- 9. The page citations will not be sorted in any record in which there are more than 200 page citations. Such records will also be flagged with three sharps. It is hard to imagine, however, that this problem would ever occur in an ordinary book index. - 50 - To search for, and correct, these flagged records, switch to Option 3, and press the <F4> function key. A menu will appear, one of whose choices is S(earch for error flags). INDEXX will search your entire file and will present each flagged record, one after the other, to be edited on-screen in the standard manner. The <F4>,S sequence can be entered in Option 3 whenever the cursor is positioned to accept keyboard data for a heading or a subheading; it is disabled when the cursor is positioned in the pages field. Entering the <F4>,S sequence erases any data pending on the screen. 3.7 Option 5 - Formatting index for printing Option 5 takes a SORTed file (from Option 4) as input and formats it for printing. It removes the three sets of quotation marks in each record, and adjusts the line lengths in the standard columnar form for indexes.[10] It also collects and indents subheadings under their common major headings. 3.7.1 Formatting options Option 5 begins by posting a series of formatting parameters, either the factory defaults or the ones previously used in the current session. You will be asked whether you want to use the posted settings. If you answer "Y", Option 5 will proceed directly to format your data file. If you answer "N", the cursor will move from one setting to the next for your response. If you wish to leave a setting unchanged, simply press <Enter>. If your file has been sorted into two lists (i.e. is a SORT-2 file), you will first be asked whether you wish four lines to separate the two lists, or whether you want the second list to begin at the top of a new page. Next you will be asked what line width you want, either `infinite' width (for which you'd enter "0") or ranging from 30 ----------- 10. If an index item exceeds the stipulated line width, that item will be split at a blankspace and continued on the next line. In the rare event that no blank occurs before the end of the physical line, the line will then be split at a hyphen. If neither a blank nor a hyphen occur, the line will be split, a hyphen will be inserted, a special error flag [###] will be appended to the end of the record, and you will be so informed on the monitor after the file has been written to disk. Only under the most extraordinary circumstances, will the "###"-flag ever occur, however. - 51 - to 72 characters per line. The `infinite' width (i.e. no line wrapping) is particularly useful if you want to read the output file (the .PRT file) into your word-processor for further formatting. If, however, you want Option 5 to produce final output, you would choose a line width in the range 30 to 72 characters. The narrower widths (30-40) allow you to print your index in columnar fashion so that you can then cut and paste the results into side-by-side columns.[11] If you do not choose `infinite' width, you will next be asked to specify the line turnover. Rather than my explaining `line turnover' in this Manual, I have programmed Option 5 to illustrate it on-screen. When you are asked for the line turnover, a window will open, illustrating what it is you are being asked for. Next you will be asked for the vertical spacing: single, double, or triple. I assume that you know what this means and that no explanation or illustration is required. When it comes time to specify the format for initials, viz. whether compact or separate, again an illustrative example will be displayed on the screen. So, too, will an example be displayed for the two types of `divider'. You will then be asked to choose between `aligned' and `run-in' style. Again, the two choices will be illustrated on screen. If you have just chosen `aligned', you will have to specify further the `indent' for the subheadings. Once again, an illustrative example will appear on the screen. Note that the indent must be less than or equal to the line turnover. (You will not be allowed to choose a value for the indent greater than the line turnover.) If the value of the indent is less than the value of the line turnover, you will be asked if you want the subheadings to be further signaled by a prefatory dash. (An example will be displayed on screen.) Some indexers and publishers particularly like this format. If, however, you have set the indent equal to the line turnover, this prefatory dash is not optional, but will ----------- 11. In creating wrapped lines of width 30-72 characters, Option 5 is smart enough to `count' only printing characters and blanks, i.e. is smart enough not to count typeface codes, e.g. the sequence "i<2001>i" would be counted as comprising four printable characters, not eight. If you were to have specified a line width of, let's say, 32 characters, Option 5, would pad such a sequence with up to 28 additional (printing) characters. - 52 - be chosen automatically (otherwise one could hardly tell the difference between the heading turnover and the start of a subheading). Once all the formatting parameters have been specified, you will be asked to confirm the list (by pressing <Enter>) or to reject it (by pressing <F10>). If the formatting parameters are accepted, Option 5 will proceed to format your data. 3.7.2 Formatted print file The output from this Option is written to disk with extension ".PRT". ".PRT" files cannot be used as input to INDEXX; they are intended to be printed with DOS COPY or PRINT; or edited and printed with your word-processor. However, since constructing a ".PRT" file leaves your ".IND" file intact on disk, you can feel free to construct a ".PRT" file anytime while an ".IND" is still in progress if you want to see what the formatted output looks like to date, or to have handy a printed list as a reminder of your entries to date. When the .PRT file has been written to disk, you'll be asked whether you'd like to View the file (V), or whether you'd like a printout now (P). You may choose either, or both. 3.8 Option 6 - Make a backup copy of index Making a backup copy of your data file, although strongly advised, is at your discretion. You will be reminded, if you've updated your file in Option 3 and have not backed it up, before you will be allowed access to Options 1, 2 or 7. If your .IND file resides in a subdirectory on your hard disk, I would strongly recommend backing it up on a floppy diskette. If your hard disk `crashes', it does you no good having your backup files on the same unreadable disk as the originals. In general, backup files should always be on physically different disks than the originals. Backup copies created by Option 6 will bear the extension ".2ND". 3.9 Option 7 - Exit to DOS If you have edited your ".IND" file but not backed it up, you will be asked if you really want to exit without making a backup. - 53 - 3.10 Review of INDEXX files ".IND" are your normal working files. The status of these files may be: EMPTY, UNSORTED, or SORTED. ".2ND" are the backups created by Option 6 of your ".IND" files above. ".OLD" are the next-to-last versions of your ".IND" files and occur in the same subdirectory as the ".IND" files. "OLD" files are normally never used: they are created purely as precautionary measures. Try not to rely on ".OLD" files. Take a moment, using Option 6, to make an updated, safer, ".2ND" file on another disk. ".OUT" are files containing all records discarded in on-screen editing sessions. They are the `trash-bin' files, or - if you prefer - the `cutting room floor'. ".PRT" are the formatted print files created by Option 5. They may not be used as input to INDEXX and are not stored in the computer. But they may be freely created while your index is still in progress if you'd like to see what your index looks like so far. "IGNORE.DEF" is the list of words created in Option 4 which are to be ignored in sorting. If Option 1 finds this file in the subdirectory designated for a new file, it will use the words in this file rather than the factory default settings for words to be ignored. 3.11 Editing INDEXX files with a word-processor WARNING: NEVER EDIT AN INDEXX .IND FILE IN A WORD-PROCESSOR WITHOUT FIRST HAVING MADE A BACKUP COPY. EDITING AN .IND FILE CARELESSLY MAY MAKE IT UNREADABLE SUBSEQUENTLY BY INDEXX. YOU WILL WANT TO HAVE YOUR ORIGINAL, READABLE, FILE AVAILABLE IF YOU SHOULD HAPPEN TO CORRUPT A COPY IRREMEDIABLY WITH YOUR WORD- PROCESSOR. All data files produced by INDEXX are in standard ASCII (text) format and may be modified - and corrected - with your word- processor. Just remember, however, you should not disturb the first twelve records in a ".IND", ".2ND" or ".OLD" file. Take especial care, too, that each data record has three pairs of matched quotation marks, and never introduce a carriage-return - 54 - (or line-feed) in a record.[12] If you do, however, disturb the quotation marks, carriage returns, etc., Option 2 will detect the error when you try to read that file back into INDEXX and you will be told exactly what needs to be repaired to get the file back into proper form. Be sure that your word-processor does not introduce formatting commands into your .IND file. Set your word-processor to produce a pure ASCII file. (In some word-processor manuals, this mode is called "text", in others "document"; unfortunately, there is no standard nomenclature.) ----------- 12. Beware of word-wrapping at the end of a line when using a word-processor to edit these files. Often word-wrapping is accompanied by the insertion of invisible carriage-returns. On certain word-processors, you can circumvent this possible source of trouble by setting the so-called `logical' line width to greatly exceed the `physical' (or monitor) line width. - 55 - Chapter 4 Supplementary program - INDXPLUS Options 8 - 11 are available in the supplementary program, INDXPLUS.EXE, also included on the distribution disk. (For the operating limits of INDXPLUS, see page 4.) To use these options, at the DOS prompt, enter: INDXPLUS. When INDXPLUS is loaded, you will be presented with a Main Menu continuing the options in INDEXX. The only valid command line options for INDXPLUS are: WHITE or BLACK or COLOR (or RGB) Unlike INDEXX, INDXPLUS produces neither .OLD, nor .2ND, files. If you want backup copies of the files produced in INDXPLUS, you will have to make them using the COPY (or XCOPY) command in DOS. INDXPLUS, like INDEXX, will check each record as it is input: for proper format, quotation marks, commas, carriage-returns, etc. If there are errors in the input files, the offending records will be displayed, so that you can correct them with your word-processor. 4.1 Option 8 - Merging files INDXPLUS offers the capability of producing a single .IND file from several segments. Thus, for example, you could devote a single file (e.g CHAP01.IND, CHAP02.IND, etc.) to each chapter of a book and then combine them (e.g. CHAP1-14.IND). You can combine any two files at a time. If you wish to combine more than two, you will first have to combine two; then combine the resulting output file with the third; then the result of this latter merger with the fourth; and so one. Of course, as you build successively larger and larger files by this means, a whole series of `stepping-stone' files will be left in the wake. If your files are on a floppy disk, you may, thus, find yourself running short of disk space, and as you proceed, may want to erase some of these spent intermediate files. Be careful, however, not to erase any of your original files. If something goes awry, you want to be able to re-start Option 8 with your original .IND files. - 56 - You will be asked to specify the paths and the filenames of the two input files, and a path and filename for the new (merged) file that is to be created. This merged file should be given a name DIFFERENT from either of the input files. All three files must bear the extension ".IND". The two files to be merged must both be sorted by Option 4 in INDEXX[13], and must both have been sorted using the identical set of sorting parameters and the identical set of `ignored' words. INDXPLUS checks the sorting parameters and the ignored words (records 9 and 11 respectively) in the file headers. If there is not precise agreement between the two files, Option 8 will abort. If the two files match properly, INDXPLUS will merge the two files, create a new header, and preserve the file parameters, etc. If INDXPLUS finds matching entries in the two files, e.g. "James, H.","","23, 88" in one and "James, H.","","55, 200" in the other, it will combine the two records, and merge and re-sort the page citations.[14] E.g. the single record which would be written to the output file would be: "James, H.","","23, 55, 88, 200" A running count will be displayed on the monitor of the records as they are being read and written. The total number of combined records will be shown when the merging is completed. As in Option 4, if there are any redundant page citations in a given record, that record will be flagged with three sharps [###] and you will be informed that there were problem cases. If the resulting output file contains no more than 2,000 data records, and if it contains fewer than (approximately) 150,000 bytes, it can be input to INDEXX for further updating in Option 3. If it exceeds either of these limits, it cannot be input to INDEXX. If the resulting (merged) file is too large to be input to INDEXX, it can still be formatted for printing by Option 9. (It is not necessary to sort this file in Option 4 first, since the output of Option 8 is always itself a sorted .IND file.) ----------- 13. See Appendix B if you are merging .IND files produced by a version of INDEXX prior to 3.00. 14. The re-sorting of page citations occurs only in those records which are combined; the page citations in the other records in the file are assumed to be sorted correctly. While INDEXX can sort only to a maximum of 200 citations per record, INDXPLUS can sort to a maximum of 1000. - 57 - 4.2 Option 9 - Formatting (oversized) .IND files Option 9 works exactly as does Option 5, but for one difference: Option 5 takes as its input a sorted file held entirely in the internal memory of the computer and is thus limited to available memory space, which cannot - because of inherent limits in the software - exceed approximately 150,000 bytes; Option 9, in contrast, takes as its input a sorted disk file whose total size is limited only by available disk space. When you choose Option 9, you will be asked for a path and a filename for the sorted file you wish formatted for printing. The file must have the extension .IND. You will then be asked for a path and filename for the .PRT file to be produced. Once you given this information, Option 9 then proceeds exactly as does Option 5. 4.3 Option 10 - Splitting oversized .IND files If you create a file, for example using the merge facility of Option 8, which is too large to input to INDEXX, you can split that file into smaller, more manageable files using Option 10. You will be asked for the path and name ("ABCDEFGH.IND") of the input file. INDXPLUS will then create names for the output files in this manner: "ABCDX001.IND", "ABCDX002.IND", etc. The sorting parameters and `ignored' words of the input file will be preserved in each of the smaller pieces. The smaller files will each be limited to 1800 records and will take up no more than 8,000 bytes of `overflow' memory (see above, p. 47). The header of each created file will cite that file's genealogy. 4.4 Option 11 - Converting old-style INDEXX files to current format Every version of INDEXX has been able to read .IND files produced by earlier versions. However, as more features have steadily been added to INDEXX, the format of files has changed. If you have .IND files created by any version of INDEXX prior to 8.xx, you will have to convert them to the format used by this current version. To convert old .IND files, simply use Option 11. It will read your old files and create new ones suitable for input to this latest version of INDEXX. - 58 - Chapter 5 Questions and answers 1. INDEXX does not load. DOS gives an error message instead: "Insufficient memory". What can I do? ANSWER: To run this professional version of INDEXX, you must have a computer with 410 kilobytes (419,840 bytes) of available RAM. The operative word here is "available". That is, you will need to have this much memory free after your computer has loaded the operating system (DOS) and whatever Terminate and Stay Resident (TSR) programs it requires, e.g. keyboard configuration, screen driver, etc. Practically speaking, these requirements mandate a computer with the maximum amount of so-called DOS memory installed, viz. 640 kilobytes. If your computer has less than this amount, you will need to install, or have installed, additional memory. If your computer does have a full 640K of memory, and INDEXX still does not load, then too much of that memory is being `eaten up' by TSR utilities loaded into memory prior to loading INDEXX. Before trying to load INDEXX, run the DOS utility CHKDSK. CHKDSK will report both the total memory installed in your computer and the available memory. If available memory is less than 420,000 bytes, INDEXX probably will not load. To run INDEXX on a computer where the TSR utilities are consuming too much memory, you will have to boot up your computer from a floppy systems-disk, i.e. one containing both DOS (i.e. the system) and a CONFIG.SYS file. The CONFIG.SYS file on the floppy disk must itself load only the minimum number of TSRs, i.e. only those absolutely necessary to configure the hardware complement in your computer. Since it is impossible for me to know which utilities you will need for your particular hardware configuration, if you do not know how to create a `bare minimum' CONFIG.SYS file, you will have to seek help from someone knowledgeable closer at hand. But be assured: the problem you are having is not due to a bug in INDEXX 8.02, just to its large size. This problem is common to a great many software applications, and many persons knowledgeable about computers will be able to lend you a hand. - 59 - (If none of the above remedies works, write to me. I could make you a custom version by removing Option 5. You could then use Option 9 in INDXPLUS instead. The custom version would not be as elegant as the factory version of INDEXX, but it would be smaller and would be serviceable.) 2. The cursor and/or some other information on the monitor is invisible. What do I do? ANSWER: Exit to DOS and re-load INDEXX (or INDXPLUS). When asked "Full-color or Monochrome monitor?", be sure to reply "M". 3. I have managed to create an .IND data file which is too large to input to INDEXX, but I need to sort this file using Option 4. What do I do? ANSWER: The overall strategy is to split this file into two (or more) smaller ones using Option 10 in INDXPLUS. Next, read each of the smaller files into INDEXX, sort them, and finally re-combine them using Option 8 in INDXPLUS. 4. I have combined two files, but find that some of the entries were not spelled exactly the same, e.g. in one file I used "wars" and in the other "war". In the resulting file, I now have two entries, "war" and "wars", where I wanted only one: "war". Do I have to combine these entries manually and re-sort the page citations manually? ANSWER: The repair is really quite easy, and certainly does not require you to sort the page citations manually. Read the file into INDEXX, and use the on-screen editing facility (see above, Section 3.5.13.2) to change the troublesome record "wars" to "war". INDEXX will then automatically search for redundancy and will combine matching records in the file, reducing the overall length of the file accordingly. You will then have to re-sort the file in Option 4. 5. INDEXX has given me the error message "Out of memory" [or "Out of string space"]. What should I do? ANSWER: Very few, if any, users will ever see such a message. If you do, your data is sufficiently unusual to have overrun the work area reserved in INDEXX. If the error message occurs while running Option 5, simply switch to Option 9. If, however, the error occurs in any other Option, you'll have to make your data files smaller. Input your .IND file to Option 10 in INDXPLUS. Option 10 will split the file into two or more smaller, more manageable, files. After editing and sorting, these - 60 - smaller files can be merged into a larger file using Option 8 and formatted for printing with Option 9. 6. Some of the records seem to have been sorted by Option 4 in strange ways. For example, the record "a priori reasoning" has ended up between "postulates" and "punishment". What's gone wrong? ANSWER: You've forgotten (as I myself sometimes do!) that this version of INDEXX ignores specified `little' words, among them "a". The sort routine in Option 4 is ignoring the `word' "a" in "a priori". The solution is simple. Either remove "a" from the list of ignored words (a solution which will affect every occurrence of the word "a" [but not, of course, the letter "a" when it occurs as part of a larger word] in your file), or alter this one record by using the on-screen editing facility in Option 3 to change "a priori reasoning" to "a priori reasoning {a priori reasoning}". In the latter case, the explicit `as-if' code will override, for that record, the `ignoring'-feature of the sort routine. - 61 - Appendix A Sorting algorithms INDEXX must sort two different kinds of data: (1) the entries in your index [fields one and two in each record], and (2) the page citations on each of those entries [field three]. The hybrid sorting method adopted for the page citations [field three] was described above pp. 38-40. The data entries pose a rather different problem. There are various ways to alphabetize strings of data. Versions 1.00-2.05 offered only word-by-word sorts (see page 48 for an explanation). Several users, however, expressed a preference for letter-by- letter sorting, and the latter method was substituted beginning with version 3.00. With version 7.00, both manners of sorting were available in Option 4. All versions since 3.00 correctly handle punctuation, foreign alphabets, data in uppercase, and lines of poetry. A.1 Pseudo-records The sorting algorithm in current versions of INDEXX begins by creating (what I call) `a pseudo-record image' of the data record. This is a record in which the first and second fields in file records are transformed in a variety of ways, including: deleting most punctuation; converting non-Greek foreign letters to English equivalents[15]; changing the case of the strings; and - in the case of letter-by-letter sorts - deleting blankspaces between words. Pseudo-records are then sorted, and - as a last step - the original records are re-ordered to correspond to the ordering of the sorted pseudo-records.[16] ----------- 15. Greek words are alphabetized so as to appear after numerical entries and before words written in Roman letters. 16. For more details on the problems associated with, and methods for dealing with, alphabetizing data, see my "Made to Order Sorts" in PC Resource, Jan. 1988. - 62 - A.2 Special cases There is, however, a very minor price to pay for this sophisticated technique. In our using proxy pseudo-records, data items which were originally dissimilar may be made to appear identical. For example, " `ideas' ","","22, 12" "ideas","","45, 678" yield identical pseudo-records: "IDEAS","","22, 12" "IDEAS","","45, 678" Thus on your final output, you may find "`ideas'" occurring before "ideas", or you may find the order the reverse. Similarly, if you have in your index an English word and a foreign word which differ only by a diacritical mark, let's say a circumflex or a dieresis, etc., the pseudo-records will be identical, and which appears first in your final output will be unpredictable. And obviously, you'd get the same kind of effect (in a SORT-1) if a capitalized word ("i<Cats>i") and a lowercase version ("cats") of that same word occurred as separate data entries. This is a minor problem. Very few users will have data which includes pairs of records which differ from one another only by quotation marks, capitalization or by diacritical marks. If, however, you must have such pairs of data in your index, by using `as-if' codes you can force Option 4 to sort the pairs in a specified order. If, for example, you wanted "`ideas'" to appear before "ideas", you could enter: "ideas {ideasb}","","23, 45" "`ideas' {ideasa}","","55, 66, 88" Or, in a SORT-1, if you wanted "i<Cats>i" to appear earlier than "cats", you could enter: "i<Cats>i {catsa}","","55" "cats {catsb}","","55" - 63 - Appendix B Updating from previous versions of INDEXX PLEASE NOTE: To input .IND files produced by a version of INDEXX prior to 8.00, you must first convert these files using Option 11 in INDXPLUS. In addition, there are enough new features in this latest version of INDEXX to make it worthwhile to work through the tutorial in Chapter 2. You will especially want to note the new, simplified, manner of entering data in Option 3. You no longer have to signal the intention to enter a subheading by typing the backslash "\" (indeed this latest version of INDEXX treats "\" as just another character, not as a special operating instruction); instead the cursor always jumps down to the subheading field after you have entered a heading. If you do not want a subheading, simply press <Enter>. The so-called Normal and String modes are now available in both the heading and the subheading fields. Extensive on-screen editing has been added to Option 3 (see pp. 43-45). Typesetting codes and `as-if' codes may be entered directly into your data file. Additional new features have been added to the sorting routine (Option 4) and the formatting routine (Option 5). Every new version of INDEXX has been able to read .IND files produced by earlier versions. However, there are some operating differences which you must understand if you are to edit .IND files in your word-processor. B.1 Changes in the design of record fields Fields in a data record were changed once with version 2.05 and once again with version 8.00. Specifically, subheadings were originally appended to the first field with a backslash, and were later placed in the second field preceded by a backslash. Beginning with version 8.00, subheadings are placed in the second field without a backslash. For example: Version 1.0: "logic \ history","","23, 78" "Russell","Bertrand","40" "Russell \ education","Bertrand","55" - 64 - Versions 2.05-7.xx: "logic","\ history","23, 78" "Russell","Bertrand","40" "Russell, Bertrand","\ education","55" Version 8.xx: "logic","history","23, 78" "Russell, Bertrand","","40" "Russell, Bertrand","education","55" B.2 Merging old .IND files Option 8 in INDXPLUS can merge only those files sorted by version 8.xx of INDEXX. If you wish to merge files produced by earlier versions of INDEXX, you will first have to convert those files to the current version, by using Option 11 in INDXPLUS, and then have to sort those files using Option 4 in INDEXX. If an old .IND file is too large to be read into INDEXX, see the special techniques for splitting and re-combining a file which are described in Chapter 5. - 65 - Index "$", use of, 20, 40 converting old .IND files, "*", use of, 16, 36 58 "\", use of (now obsolete), copying the distribution 64 diskette, 7 "###", use of, 46, 50, 51, credits, 6 57 cross-references, 37 "000", use of, 19, 38 within (sub)headings, 19, "< >", use of, 17, 36 38 "{ }", use of, 18, 37 within page citations, 38, <Arrow>-keys, 42 40 ASCII, 3, 37, 38, 54 <Ctrl>-key combinations, 42 listing (nos. 32-126), 38 automatic saving to disk, data records, 33, 42, 54 13, 22 number of, 47, 57 frequency of, 27 default settings Automode, 12, 41 changing, 26, 27 start-up, 30 <Backspace>-key, 42 <Del>-key, 42 backup copies, 24, 53, 56 delimiters batch file, 30 between fields, 33 block (of records), 44 in command line, 30 boldface, 36 disclaimer, 6 DOS, 3, 29, 53, 56 canceling an Option, 26 CHKDSK, 59 CapsLock, 10 CONFIG.SYS, 59 Chicago Manual of Style, 7 MODE, 26 color, 27 foreground/background, 27, editing 28 file record, 43 graphics adapter, 8, 26 input line, 43 primary/secondary, 27 on-screen, 2, 21 update speed, 28 EDLIN, 30 combining files, 4, 31, 56, <End>-key, 42 60 error checking, 56 combining records, 45, 50, error flags 57 ###, 46, 50, 51, 57 command line error messages, 15, 22, 31, INDEXX, 29 47, 54 INDXPLUS, 56 "... in Module...", 5 commas, 33 "Insufficient memory", 59 surrogate, 20, 39, 40 "Only 2,000 entries confirmation permitted", 47 lengthened entry, 35 "Out of memory", 47, 60 new item, 10, 28, 34 "Out of string space", 60 partial match, 14, 28, 34 errors, reporting, 5 - 66 - <Esc>-key, 34, 42 <F5>, 12, 41 <F7>, 12, 41 fields, 12, 33, 64 <F8>, 42 markers between, 12, 33 <F9>, 16 markers within, 12, 40 <F10>, 22, 26, 34 file status review of usage, 46 EMPTY, 10 fuzzy logic, 3, 16, 34 SORT-1, 48 SORT-2, 48, 51 given name(s), 2, 11, 14, SORTED, 25 32, 33, 34 UNSORTED, 10, 25 files help window, 42 .2ND, 24, 27, 31, 53, 54, <Home>-key, 42 56 hot keys, 9 .BAT, 30 .IND, 24, 27, 31, 33, 53, <Ins>-key, 42 54, 56, 57, 58 insert mode, 42 .OLD, 27, 54, 56 installing INDEXX, 7 .OUT, 45, 54 italics, 36 .PRT, 53, 54 data, 9 line width, 51, 54 extensions, 31 loading INDEXX, 8 headers, 12, 49, 50, 54, 57 Main Menu, 9, 26, 56 IGNORE.DEF, 49, 54 Manual of Style, 7 INDEXX.EXE, 4, 7 Manual, printing the, 7 INDXPLUS.EXE, 4, 7, 56 matches, 14, 57 MANUAL.BAT, 4, 7 case sensitive, 35 MANUAL.DOC, 4, 7 partial, 3, 16, 34 ORDRFORM.DOC, 4 definition of, 34 PAYMENT.DOC, 4, 5 perfect, 11, 16 READ.ME, 4 memory available, 47 review of INDEXX files, 54 monitor trash bin file, 45 invisible characters, 60 foreign alphabets, 37, 62 monochromatic adapter, 8 diacritical marks, 63 multiple-word (sub)headings, formatting, 3, 34, 51, 58 32 aligned/run-in, 52 divider, 52 Names/Subject index, 48 indent (for subheadings), normal data entry mode, 15, 52 16, 28, 32, 33 line turnover, 52 line width, 51 operating limits, 3, 47, 50, Names/Subjects, 51 57, 58 on-screen examples, 52 operating requirements, 3, persons' initials, 52 59 subheading prefix, 52 Option "0", 26, 29 vertical spacing, 52 Option "1", 9, 30 FREQ, 29 Option "2", 24, 31, 54 function keys Option "3", 9, 25, 27, 32, <F1>, 15, 28, 32 57 <F2>, 11, 17, 36 Option "4", 20, 22, 37, 48, <F3>, 12, 13, 41 57, 60, 62 <F4>, 21, 43 - 67 - Option "5", 22, 24, 34, 51, character two hundred and 58, 60 twenty-five, 49 Option "6", 53 float/sink, 48 Option "7", 53 ignoring `little' words, Option "8", 31, 47, 56, 60, 22, 49, 61 65 incorrect order, 61 Option "9", 56, 58, 60, 61 letter-by-letter, 48, 62 Option "10", 31, 56, 58, 60 oversized files, 60 Option "11", 56, 58 page citations, 50 oversized files, 58, 60 SORT-1, 48 overwrite mode, 42 SORT-2, 48 word-by-word, 48, 62 page citations, 3, 13, 38, sound, 27 47, 57, 60, 62 splitting files, 58, 60 duplicates, 46 string data entry mode, 15, redundancy in, 50, 57 28, 32, 33 path (DOS), 9, 24, 29 subfield, 12, 20, 40 persons' initials, 34, 52 subheadings, 2, 33 persons' names, 32, 33 surname(s), 2, 11, 14, 32, <PgUp> / <PgDn>-keys, 42 33, 34 poetry, 62 printer, 24 trash bin file, 45 printing the formatted file, TSR (Terminate and Stay 22, 34 Resident) utilities, 59 prompts, 10 tutorial, 7 pseudo-record, 62 typefaces, 3, 17, 36, 39, 52 questions and answers, 59 underlining, 36 quotation marks, 33, 51, 54 updating from previous versions of INDEXX, 1, re-ordering of given names 64 and surnames, 32 upper- and lowercase, 62 record, 12 user's fee, 5 pretend' record, 44 retrieval of record for viewing the data file, 11 editing, 21 viewing the formatted file, Roman numerals, 19 22 single-word entries, 32 wildcard "*", 16, 36, 39 sorting, 2, 38, 57, 60, 62 word-processor, 12, 18, 30, `as-if', 18, 37, 39, 40, 31, 53, 54, 64 49, 61, 63 word-wrapping, 43, 54 - 68 -