home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Crawly Crypt Collection 1
/
crawlyvol1.bin
/
apps
/
database
/
refbase
/
refbas14.doc
< prev
next >
Wrap
Text File
|
1992-08-17
|
66KB
|
1,284 lines
RRRRRRR EEEEEEEE FFFFFFFF BBBBBBB AAAAAA SSSSSS EEEEEEEE
RR RR EE FF BB BB AA AA SS SS EE
RR RR EE FF BB BB AA AA SS EE
RRRRRRR EEEEEE FFFFFF BBBBBBB AAAAAAAA SSSSSS EEEEEE
RR RR EE FF BB BB AA AA SS EE
RR RR EE FF BB BB AA AA SS SS EE
RR RR EEEEEEEE FF BBBBBBB AA AA SSSSSS EEEEEEEE
Manual to REFBASE Version 1.4 - 14 August 1992
Author: Jan M. de Boer
Department of Nematology
Wageningen Agricultural University
P.O. Box 8123
6700 ES Wageningen
THE NETHERLANDS
Telephone: 08370-82427
Fax: 08370-84254
Version 1.4 of REFBASE succeeds earlier versions which were released on
the following dates:
23 September 1991 - Preliminary version
12 January 1992 - Version 1.0
11 February 1992 - Version 1.1
13 April 1992 - Version 1.2
9 July 1992 - Version 1.3
This manual succeeds the manual of version 1.0, and it includes the
changes that have been made in the versions 1.1 to 1.4. For convenience a
separate file "UPDATE.DOC" is included as well to inform earlier users of
the changes and repairs that have been made.
REFBASE 1.4 is Public Domain. You are allowed to pass on copies of the
software provided that all original files and programs of the package are
included.
The author has been using REFBASE intensively for several months now,
and version 1.4 appears to be free of major bugs. Nevertheless, the
author takes no responsibility for any damage to data or hardware caused
by remaining defects in the program.
REFBASE was written with the GFA-Basic Interpreter V. 3.07, and compiled
with the GFA-Basic Compiler V. 3.03. A few pieces of assembly code have
been included as well. These were compiled with the GFA-Assembler.
This manual contains the following sections:
INTRODUCTION
START/QUIT
MAIN MENU
INPUT
DISPLAY
FONTS
AMEND RECORDS
BLOCK FUNCTIONS
SCREEN COLOUR
SORTING
GOTO
SEARCHING/SELECTION
TEXT MODE
TAGGING
LOADING FILES
MEMORY MANAGEMENT
SAVING FILES
DELETING RECORDS
FIND DUPLICATES
CASE CONVERSION
RECORDS MENU
FIELDS MENU
MODE MENU
THE MOUSE
KEYWORDS
PRINTING
CHAIN SEARCH
FILE FORMAT / DATA EXCHANGE
SUPPORTING PROGRAMS
LIMITS OF REFBASE
---------------------------------------------------------------------
INTRODUCTION
REFBASE is a database program for storing scientific references. REFBASE
was designed primarily for handling titles from publications which have
appeared in journals; for storing book-titles the record layout is
probably somewhat less suitable.
REFBASE loads its data files completely into the memory of the computer.
The maximum size of files that can be displayed and sorted is therefore
limited by the amount of RAM that you have installed. If you have more
records than will fit in the computer's memory, you can store them in
several smaller files, which can then still be "chain searched" in a
single run.
The basic design of REFBASE was taken from Data Manager Professional (the
former TRIM-base), and many key functions are therefore still identical
in both programs. The aim was to develop a program that would overcome a
number of severe limitations of DM-Professional, and yet offer the same
user-friendly interface. Compared with DM-Professional, in REFBASE a
number of features have been greatly improved, such as (1) fast loading
of multiple files in memory, (2) fast scrolling, (3) easy splitting and
merging of files, (4) handling of large text fields for storing
abstracts, (5) high speed export of fields for keyword extraction, and
(6) instant printing of neat reports without having to switch to other
program modules and search for the correct report forms. In addition many
new features have been added.
REFBASE was developed on a Mega ST4 with TOS 1.4. Although REFBASE will
run on an ST with 1 Mb of memory and a single sided floppy disk drive, it
is advised to use at least 2 Mb of RAM and a hard disk. REFBASE will only
run in ST high resolution screen mode (640 x 400 pixels). It is advised
to install a screen accelerator such as Turbo ST or NVDI; this will
significantly speed up scrolling through the records.
REFBASE has not yet been seriously tested on the Atari TT. I think it
will probably run without problems on this machine as well, provided that
the program is started in ST-RAM. If you start REFBASE in TT-RAM problems
with the screen may be expected and possibly the program crashes right
from the start.
-------------------------------------------------------------------
START/QUIT
After starting REFBASE the load menu is displayed. To load a file see
LOADING FILES. Click on EXIT or press ESC to enter the main menu. You can
quit REFBASE from the main menu by clicking on EXIT, or pressing ESC, and
answering the alert box. With CONTROL-X you can quit REFBASE immediately
from most points in the program and return to the Desktop without having
to answer an alert box.
--------------------------------------------------------------------
MAIN MENU
Pressing HELP in the main menu gives a summary of the key functions.
The two rows of menu options at the top of the screen are selectable both
via the function keys and with the mouse.
With F1 (Acc/Cpx) a drop down menu is created with which you can access
accessories such as the CPX control panel. You return to the main menu
by pressing ESCAPE or clicking on "Back to main menu". (During Display
the function of F1 is changed to a Fix/Float cursor toggle)
The other function key options are explained in the different sections of
the manual.
Other mouse selectable functions in the main menu are:
Chain search - See Chain search
Display all records - See Display
Display selected records - See Display
Display tagged records - See Display
Show selection rule - See Searching/selection
Save mode preferences - See Mode menu
Read drive - to show/enter the drive from which files will
be loaded
Write drive - to show/enter the drive to which files will
be written
EXIT - to return to the Desktop
Disk space - To ask for the free space on a specified
drive
File size - To calculate the size of the file(s) loaded,
and the average record size
Not selectable items are:
The date and clock - change them via the control panel
Free memory - this value is automatically updated whenever
operations affecting memory space are
performed.
The file name "window" - this only shows the file(s) currently loaded
--------------------------------------------------------------------
INPUT
Click/press on INPUT (F5) to enter new records. Each record consists of
10 fields for the following parts of a reference:
FIELD LABEL SIZE in USE
characters
1 authors 70 x 2
2 year 4
3 subject 8 For dividing records in main categories
4 read/ 8 To indicate whether or when an article
document has been read; you can also use this field
to store the document type (e.g. article,
note, review)
5 file 6 For a short code to indicate the file from
which record originates, or to which it
must be added.
6 title 70 x 5
7 journal 50 including volume and page numbers
8 date 8 To store the date you entered the record
9 keywords 70 x 6
10 remarks 70 x 21 A multi-functional large text field,
suitable for entering remarks, storing
author's addresses, abstracts, or
marking labels.
The size and position of these fields is fixed. It is not possible to
rearrange or modify them to your own demands. All fields are text fields.
You start entering data in the first field. With the INSERT or CURSOR
DOWN keys you can move to the next field, and with the CURSOR UP key you
can return to previous fields. You can also place the cursor at the start
of any field by clicking on it with the mouse.
The fields 1, 6, 9, and 10 are "block" fields consisting of multiple
lines. Usually text in a block field is automatically formatted when
entering or inserting text. If this does not occur, you can induce a
reformat by deleting (and re-typing) a character in the preceding line.
The block-editor follows the style of 1st Word Plus, i.e. the lines in a
block field (=paragraph) should end with either a space of a hyphen (-).
Internally the lines of a block field are always stored as a single line;
it is therefore not possible to force line separations in a block field
by pressing the RETURN key.
As soon as you enter the remarks field with the cursor it will be shown
in full size. You leave the remarks field again with the CURSOR UP key.
The field sizes shown in the table above are the limits for manual data
entry. If you wish you can overrule these limits with the "paste block"
or "search and replace" functions. The remarks field actually can display
42 lines (two screens) of text, and is therefore suitable for storing
abstracts.
Other key functions are:
BACKSPACE - erase character left of cursor
DELETE - erase character at cursor position
RETURN - moves cursor to next line.
CLRHOME - erase the contents of a field
CONTROL-B - show block
CONTROL-P - paste block at cursor position
CONTROL-K - enter keywords menu
SHIFT + CURSOR LEFT/RIGHT - moves cursor to start/end of line
SHIFT + CURSOR UP/DOWN - moves cursor to first/last line of field
CONTROL + CURSOR LEFT/RIGHT - moves cursor one word left/right
CONTROL + DELETE - deletes word at cursor position
UNDO - restores field contents
There are no provisions for entering special characters such as é, Ä, or
µ. With the new TOS versions 2.06 and 3.06 ALTERNATE + ASCII code will
automatically allow the entry such characters.
The contents of field 5 is automatically converted to upper case after
entry.
In fields 4 and 8, pressing the '#' key enters the system date in the
format YY/MM/DD (this format is easy to sort).
The length of the journal field has deliberately been kept short (50
characters) so that it can be easily displayed in table view. The journal
field is therefore only suitable for storing abbreviated journal names.
To ensure a correct sort, it is advised to be very consequent when typing
in the journal field. The author uses the following format:
J Examples 012: 12-18 (2)
The volume numbers consist of three digits, if necessary one or two of
them are typed as zero's. The starting page number always occupies four
character positions, of which up to three of may be filled up with
spaces. If desired, the issue number can be placed at the end of the line
in brackets; in this way it will have no influence on sorting.
By pressing TAB the record is accepted and added to the file.
Automatically a new blank record is displayed. You can stop data entry,
or reject a record, by pressing ESCAPE.
During data entry a beep will sound when CAPS LOCK is on.
After entering new or amended records by pressing TAB, (1) all fields
(except the journal field) are automatically checked for double spaces,
which can be removed by answering an alert box, (2) in block fields,
lines that did not yet end with a space or a hyphen "-" are automatically
provided with a terminal space, and (3) any lingering spaces at the end
of the fields are automatically removed.
---------------------------------------------------------------------
DISPLAY
Press "A" or "F3", or click on "Display" or "Display all records" to show
the records in the file(s) currently loaded. Alternatively, pressing "S"
or "D" or clicking on "Display selected records" or "Display tagged
records" will respectively show the selected or the tagged records.
You can scroll through the file with the CURSOR UP and CURSOR DOWN keys.
By simultaneously pressing CONTROL or SHIFT, you can scroll through the
file in steps of 20 records or 1/10 file length.
During display, the keyboard sound is switched off, so that you will not
continuously hear the cursor keys while scrolling.
With F8 ("View") or INSERT you can switch between "Card view" and "Table
view". In "Table view" you can use F9 ("Expand") to switch between a
compact one-screen table, and an expanded five screen wide table. In the
expanded table mode you switch between screens with the CURSOR LEFT and
CURSOR RIGHT keys.
In "table view" pressing F1 switches between a fixed and a floating
cursor. The mouse can be used to place the cursor on any desired record.
In "table view" pressing CONTROL CURSOR UP/DOWN will jump one screen
full of records backward/forward. You can also scroll one page forward
or backward by using the SPACEBAR or the BACKSPACE key.
From the "Table view" you can press F10 to enter the "Title view" mode.
You can use the CURSOR UP/DOWN keys in combination with the CONTROL or
SHIFT key to scroll through the titles. Pressing CURSOR LEFT or CURSOR
RIGHT will start an automatic soft scroll, which can be stopped by
pressing any key. In "Title view" the "expand" function (F9) is inactive.
You can switch between "Title view" and "Card view" by pressing F8 or
INSERT. The normal table view is restored by pressing F10 again.
In the mode menu the speed of the automatic "Title view" soft scroll can
be varied by choosing between 0 and 9 "units of pause" between
consecutive records (one unit of pause is 400 ms). You can interrupt the
soft scroll also DURING a pause period by pressing any key.
You can tag a record by pressing "T". The tagged record is now shown in
grey. To release a tag press "T" a second time. You can use the tagging
facility to pick out manually a selection of records, for instance for
printing, or saving in a separate file. (see also TAGGING).
With CONTROL-CLRHOME you can jump to the first record in the file; with
CONTROL-Z you can jump to the last record in the file.
The numeric key pad can also be used to scroll through the records:
+-------------+-------------+-------------+
| (7) | (8) | (9) |
| Home | Cursor Up | Page Up |
+-------------+-------------+-------------+
| (4) | | (6) |
| Cursor Left | | Cursor Right|
+-------------+-------------+-------------+
| (1) | (2) | (3) |
| End | Cursor Down | Page Down |
+-------------+-------------+-------------+
In "Card view" a copy of the current record can be made by pressing
CONTROL-D. The duplicate record will then be placed at the bottom of the
file.
In "Card View" pressing CONTROL-P prints the current record in full
"card" format to the parallel printer port, with a maximum line length
of 78 characters (Pica width). No printer control codes are given, except
for a formfeed at the end of the record.
In "Card view" pressing CONTROL-M activates a menu with which marking
labels can be inserted at the start of the Remarks field. Use this to
indicate what you want to do with the record. The marking labels are
placed between {}-brackets. You add (or remove) the labels by pressing
their first letter. Press "O" (= Order library) if you want to order a
copy of the article via the library, or "R" if you intend to ask for a
reprint. By pressing "D" (= Done + date) you can mark the date that the
request for the article was sent away. Alternatively use "F" if you want
to fetch the article from the nearest library, or "L" if you just want to
look it up. If by any chance a journal issue was not present in the
library (e.g. to the binder) you can come back later, and mark the record
with "M" (= Missing + date). With "A" (= Add) you can enter a text
yourself. Pressing "C" (= Clear) removes all marking labels in the
current record. You can switch off the mark menu by pressing CONTROL-M a
second time.
Both in card view and table view pressing HELP displays the remarks field
in full screen. In order to accommodate for large abstracts to be stored
in this field, the remarks field actually consists of two full screens,
and you can switch between them by pressing the CURSOR UP and CURSOR DOWN
keys. This two-screen facility, however, will only work if there is
enough text in the remarks field to fill more than one screen. Also it
does not work during manual data entry: in that case you are allowed to
type in only a single screen full of text. You can return to the normal
display mode by pressing HELP, ESC, or UNDO.
If the text in a block field is too large to be displayed, the bottom
line of the field will terminate with a ">>". If such a field occurs,
editing its text in the normal way with "amend" (F6) may give problems
because the limit for entering text in this field is already reached.
With a small block field (Authors, Title, or Keywords) it is possible to
copy its contents to the block buffer (see below), which offers more
space for editing. Then, after editing, you can copy the text back to the
field.
In "card view" clicking with the mouse on a field activates a menu with
which its contents can be erased, copied to the block buffer, or modified
by paste operations (see BLOCK FUNCTIONS).
You leave the Display mode by pressing or clicking F3 again, or by
pressing ESC or UNDO.
--------------------------------------------------------------------
FONTS
During record display in "table view" and during display of keywords
pressing "F" or "f" switches between the 8*16 and 8*8 font. In "card
view" and "title view" only the 8*16 font is used.
------------------------------------------------------------------
AMEND RECORDS
In "Display mode" you can modify the contents of the current record by
pressing or clicking F6 ("Amend"). The cursor is placed in the first
field, and you can add or delete text as described under INPUT. Press TAB
to accept the modifications, press ESCAPE to exit "Amend" and leave the
original record unchanged.
--------------------------------------------------------------------
BLOCK FUNCTIONS
You can cut and paste text between records using the block buffer.
Although in theory the maximum size of this buffer is 32767 characters,
the block editor will only allow you to type in a single screen full of
text.
You enter the block editor by pressing "B" in the main menu or during
display. If you are entering or editing a record, access to the block
editor is denied, and you can only "show" the block by pressing CONTROL-
B. After having typed in a block text you must press TAB to save it to
the block buffer. Pressing ESC rejects the entered block text, and leaves
the original block unchanged. When records are displayed in "card view",
clicking with the mouse on any field activates a selection box with which
you can copy the contents of this field to the block buffer, or paste the
block contents to the start or the end of the field. "Fill with block"
completely replaces the field contents by the block text. Block copies
that are made like this ignore any limitations imposed on the field size.
If you are entering or editing records, pressing CONTROL-P will paste the
contents of the block buffer at the current cursor position. In this case
the block copy will stop as soon as the field is full.
In the Fields menu, choosing "Paste block" will allow you to insert a
block text in a given field in a range of records.
--------------------------------------------------------------------
SCREEN COLOUR
In the main menu and during "Display" the screen colours can be inverted
by pressing "I" or CONTROL-I. Inverting the screen influences the line
length in "Table view": with white characters on a dark background, the
full width of the screen is used, and two extra characters per line are
displayed. The menu bar at the top of the screen will always be shown in
black.
--------------------------------------------------------------------
SORTING
The records can be sorted by pressing or clicking on F4 ("Sort") and
selecting the desired sort field in the selection box. The current sort
is marked with an asterisk (*). Via the "Mode" menu you can indicate by
activation (=>) of "Recall previous sorts" whether or not you want
REFBASE to store the sort sequence. If this option is active, then
"stored sorts" are marked in the sort selection box with a dot (.). If
you sort on a field marked with a dot, REFBASE does not perform a sort,
but instead uses the stored sort sequence to arrange the records. With
large files this option allows instant switching between different sort
sequences (e.g. authors and journals) without having to wait. The maximum
number of stored sort sequences is three. If more new sort sequences are
added, the oldest ones are discarded. Stored sorts are erased if extra
files are loaded, or if records are added or deleted.
For multi-level sorts you simply do multiple sorts after each other,
starting with the innermost level. You may need to switch off "Recall
previous sorts" via the Mode menu if you want to sort like this.
During a sort only the first 60 characters of a field are considered.
The sorting algorithm uses a recursive "quicksort" procedure. Generally
this sort will be done very quickly. However, there is an exception: if
two identical or nearly identical files are loaded that were already
sorted on a given field, then a first sort on this same field may take
much more time than usual.
--------------------------------------------------------------------
GOTO
By pressing or clicking on F7 ("Goto") during record display you can
instantly jump (by binary search) to a specified position in the file.
"Goto" only works with the field that the records are sorted on.
Depending on the display mode, you can jump through all records in the
file, or only through a subset of selected or tagged records.
--------------------------------------------------------------------
SEARCHING/SELECTION
By pressing or clicking on F2 ("Select") you can search for records using
one or more selection criteria. By answering selection boxes you can
choose successively a field to be searched, a condition, and a search
text. In addition to the normal fields, it is possible to search the
"BASIC" pseudo field, which is a simultaneous search in both the title
and the keywords field.
A maximum of 5 selection criteria can be combined by clicking/pressing F3
("AND") or F4 ("OR"). As soon as the selection rules have been entered,
the search can be started by clicking/pressing F10 ("Accept"). After the
search, an alert box shows the number of records found and asks whether
you want to see them.
After each search the record selection remains available. The selected
records can be displayed at any time by clicking on "Display selected
records" or pressing "S". Pressing "R" or clicking on "Display selection
rule" shows the selection rule that belongs to the current record
selection.
When searching with "Includes" and "Does not include", the "?" and "*"
may be used as wildcards. The "?" is a placeholder for one unknown
character. Similarly "???" specifies three unknown characters. The "*" is
placeholder for a variable number of unknown characters. It is not
allowed to let the search text begin or end with a "*" or "?", nor is it
allowed to use the combinations "?*", "*?" or "**". If desired, searching
with wildcards can be switched off via the "Mode" menu.
"Includes" and "Does not include" searching uses a "quickfind" algorithm,
which was described in the Dutch magazine "ST", issue 31 (may/june 1991).
Although this search routine appears to function correctly, it still is
possible to switch it off in the "Mode" menu. Searching will then proceed
with the standard search command (= INSTR) of GFA-BASIC. The speed of the
quickfind algorithm depends on the length of the search text: the more
characters you type in, the faster the search will be.
The quickfind algorithm works both in "ASCII" and "Alpha" text mode (see
below). The alphabetical search, however, will only work correctly with
the normal character set (ASCII code <128). If you want to search in
"Alpha" mode for special characters such as ä and Ä you must inactivate
the quickfind routine via the mode menu.
Via the "Mode" menu you can choose for searching with what I have called
"partial equation". This implies that in tests involving "<=", "=", ">",
or ">=" only the first few characters of a field are examined. The effect
is that searching is done in a more 'natural' way, instead of following
strict computer logic. For instance the search condition " Authors <= E
" will normally not select records of which the authors field begins with
"Ea.." to "Ez..". With partial equation these records are included as
well, so that you end up with all authors of which the name begins with
A, B, C, D, or E, just as you wanted. If you deselect 'partial equation',
the search conditions "Begins with" and "Does not begin with" are changed
to "Is" and "Is not".
Searching in the BASIC pseudo field (Title plus Keywords) can only be
done with "Includes" and "Does not include".
Most of the functions of REFBASE can be applied to the selected records,
such as saving, printing, deleting, etc.
If you press CONTROL the F2 menu text will change to "Repeat". Clicking
on "Repeat" or pressing CONTROL-F2 allows you to repeat a search with the
latest selection rule.
---------------------------------------------------------------------
TEXT MODE
After starting REFBASE the default settings are to search and sort
alphabetically, i.e. no distinction is made between upper case and lower
case characters. By clicking on "Alpha" or pressing SHIFT-F5 you can
activate the "ASCII" search mode, in which upper and lower case
characters are treated differently during a "Select" operation. (This
switch is also available in the Chain Search and Keywords menus; see
below). If you want to sort your records by ASCII-code, you must activate
this option separately via the "Mode" menu. This automatically implies
that a number of other options ("GOTO", and "Find Duplicates") also will
distinguish between upper and lower case. "Search and Replace" (see
below) only works in ASCII mode.
---------------------------------------------------------------------
TAGGING
In "Display" mode records can be tagged by pressing "T". The tagged
records are marked with a "grey colour". From the main menu, you can
display the tagged records by pressing "D" or clicking on "Display tagged
records". To release a record tag simply press "T" again.
With "TagSel" (SHIFT-F9) you can tag all currently selected records. and
with "ClrTags" (SHIFT-F10) all current tags are erased. In addition the
"Records" menu has the options "Untag selected records", "Invert tags",
and "Tag all records".
Most of the functions of REFBASE can be applied to the selection of
tagged records, such as saving, printing, deleting, etc.
---------------------------------------------------------------------
LOADING FILES
By clicking on "Open" (SHIFT-F1) you enter the load menu, and the
directory of the current drive is searched for REFBASE files ending with
".REF". You can also ask to display all files by clicking on the box with
" * ". Use the mouse to select one or more files to be loaded. By keeping
the mouse button depressed, you can select multiple files at once. Also
you can select them all, by clicking on "ALL". You can cancel a selection
by clicking on the file while keeping the CONTROL key depressed. UNDO
deselects all files. To load a selection of files click on LOAD. If you
already have loaded a file this will be erased and replaced by the new
file(s). Alternatively, you can choose to load extra files by clicking on
MERGE.
If multiple files are loaded, each new file will be placed at the
"bottom" of the previously loaded file(s), and all loaded files will be
treated as a single large file. The "window" in the main menu shows which
files are loaded. After each load or merge operation any stored sort
orders are erased, and the GOTO facility becomes inactive until you have
sorted the file again. You can interrupt the loading of files by pressing
ESC.
Folders are marked with a "*" before their name. Open them by clicking on
their name. To close a folder, click in the box showing the drive path.
You can switch between drives by clicking on the drive buttons. Drives
"C" to "P" are also selectable via the keyboard by pressing the
corresponding drive letter. If you select drive B, for security you will
be asked once whether this drive is really switched on. You can also
specify the drive for loading files in the main menu, by clicking on
"Read drive" and typing in the drive letter.
Read only files are marked with a "°" before their name.
Hidden files are marked with a "#" before their name.
The filename extension button ".ISS" is meant for issues files which were
created with CC_HELP.PRG. By clicking on "???" you can enter a directory
file-extension yourself. This extension will then become selectable under
the left most "extension button".
In the load menu a maximum of 68 files and folders can be displayed. If
the number of files/folders exceeds 34, their names will automatically be
printed in small text. If a directory contains more than 68 files or
folders that match a given search extension, the reading of this
directory is stopped and a warning is given.
Finally, the load menu offers some extra options:
* With "NEW" you can erase the current file(s) from memory and start
typing in a new file (this is equivalent to quitting and starting
REFBASE)
* With "DEL" you can delete a selection of files, provided that
they are not write-protected.
* With "RENAME" you can rename a series of selected files. One by one
you can edit their names. Only the letters A to Z, the dot (.) and
the underscore (_) are accepted as valid characters. To skip a name,
simply press RETURN, or enter an empty name.
* With "ATTR" you can modify the file attributes of the selected
files. "Hiding" a file makes it invisible in the normal GEM-Desktop.
For the rest, hidden files can be read normally by REFBASE.
For NeoDesk users, multiple files can be loaded automatically from the
Desktop by dragging them on the REFBASE icon. The maximum number of files
that can be loaded in this way is limited by the amount of space that
GEMDOS reserves for storing the file names with their drive paths: their
total length should not exceed 127 characters. You can only load .REF or
.ISS files. A mixture of .REF and .ISS files is not allowed: if the first
file in the sequence is a .REF file, any following .ISS files will be
ignored, and vice versa. (Attempts to get a similar autoload facility
working correctly with the "Install Application" option of the TOS 1.4
desktop have failed until now).
REFBASE assumes that the files to be loaded, either in the load menu, or
in the keywords menu, are text files. If you try to load any other type
of file errors may occur, and garbage will be displayed on the screen.
You can leave the load menu by pressing ESCAPE or clicking on EXIT.
---------------------------------------------------------------------
MEMORY MANAGEMENT
The maximum number of records that can be loaded in REFBASE is shown by
the "Space" indicator in the main menu, directly after starting the
program. If you load a file, the number of loaded records is shown by
"Loaded", and "Space" will show the record space left. If during a load
operation the maximum number of records is reached, loading will be
stopped.
REFBASE adjusts the amount of reserved record space to the amount of RAM
available in the computer. In doing so, the program assumes an average
record size of 400 characters (=bytes). This is the typical size of
records with a reasonable amount of keywords and without abstracts.
Therefore the maximum number of records that REFBASE will load is about
600 with 1 Mb of RAM, 2300 with 2 Mb RAM, and 5800 with 4 Mb RAM.
Records with abstracts consume much more memory space (up to 3000 bytes
per record). If such records are loaded, the "space" indicator is
unreliable, and will over-estimate the record space. For this reason
REFBASE will also check if the size of files to be loaded or merged does
not exceed the available amount of RAM, and a warning will be given if
this is the case. This check for sufficient RAM will also include the
size of any keywords files already loaded or deleted records still stored
in memory, and you may need to erase a large keyword file or do a garbage
removal first, before you can load a large data-file.
You can ask for the average record size in a file by clicking on "File
size" in the main menu.
---------------------------------------------------------------------
SAVING FILES
Click on "Save" (SHIFT-F2) to save records to disk. If selected or tagged
records are available, you can choose to save only these records. If you
save the records to a file that already exists, a warning will be given,
and you will be offered three options: overwrite the file ("replace"),
add the records to the end of file ("append") or stop the save operation
("cancel"). If you append records to a file on disk, be sure that this
file does not become too large for loading by REFBASE. If this should
happen anyway, you can use REFSPLIT.PRG to recover the file (see
SUPPORTING PROGRAMS)
You can specify the drive for saving record files in the main menu, by
clicking on "Write drive" and typing in the drive letter. This drive
selection is also valid for other save operations such as export, or
printing to disk.
During all save operations there is NO CHECK FOR SUFFICIENT DISK SPACE.
You can do this yourself by clicking on "Disk space" in the main menu,
and typing in the letter of the drive. Clicking of "File size" will give
the size of the file that you want to save.
During a save operation, the normal screen disappears, and a "snowstorm"
appears. This is the file, which is being written to disk in blocks of
nearly one screen full.
(If a save operation ends with a block of about one fifth of a screen
full of "snow", then for unknown reasons REFBASE pauses for a moment,
before writing the information to disk. This is probably a "bug" in GFA-
Basic, for which I know no remedy).
---------------------------------------------------------------------
DELETING RECORDS
Records can be deleted manually by pressing DELETE. Also it is possible
to delete selected or tagged records via the "Records" menu. Deleted
records are not erased from memory: they keep occupying space until a new
file is loaded. If you have deleted a large number of records, you can
permanently erase them from memory by choosing the "Garbage removal" in
the "Records" menu. The space which they occupied then becomes available
again, for instance for loading extra files with MERGE.
Use "Undo record deletions" in the "Records" menu to retrieve all deleted
records or the latest deleted record. These records will then be placed
at the bottom of the file, and they are accessible by displaying all
records. If by any chance a deleted record was tagged or selected, this
will no longer be the case.
---------------------------------------------------------------------
FIND DUPLICATES
After sorting on the "Journal" field you can look for duplicate records
by choosing "Find duplicates" in the "Records" menu. REFBASE will then
check if there are references of which the journal fields are EXACTLY the
same, in which case they are scored as "duplicates". The duplicate
records are stored as "selected records", which can be displayed by
pressing "S", or clicking on "Display selected records". Also, of each
set of N duplicate records, N-1 are tagged. In this way deleting selected
records will delete all duplicates, whereas deleting tagged records will
retain one copy of each duplicate record in the file. Searching for
duplicates will erase all current record tags.
Note that this duplicate finding routine will fail if two or more
different references occur on the same page of the journal (e.g.
abstracts from symposia, or book reviews).
--------------------------------------------------------------------
CASE CONVERSION
In the "Records" menu the references can be converted to upper case or
lower case by clicking on "case conversion". During the case conversions
the special characters "ÇÄÅÉÆÖÜÑØŒÀÃÕIJ" are changed to "çäåéæöüñøœàãõij"
and vice versa. The "Fields" menu also offers the possibility of case
conversion, with the difference that this time the conversion can be
applied to a single field.
---------------------------------------------------------------------
RECORDS MENU
The options of the Records menu have already been discussed in the other
sections of the manual:
"Untag selected records" see TAGGING
"Invert tags" see TAGGING
"Tag all records" see TAGGING
"Delete selected records" see DELETING RECORDS
"Delete tagged records" see DELETING RECORDS
"Undo record deletions" see DELETING RECORDS
"Find duplicates" see FIND DUPLICATES
"Case conversion" see CASE CONVERSION
"Garbage removal" see DELETING RECORDS & MEMORY MANAGEMENT
---------------------------------------------------------------------
FIELDS MENU
"Fields" menu has the following options, which can be applied to either
all records, or a selection of records:
With "Export field" the contents of a single field can be written to a
file on disk. The most useful application is exporting the keywords field
for generating a keyword list with KEYWORDS.PRG.
With "Search and Replace" you can search for a text string in a field,
and replace it by another text string. "Search and Replace" only works in
ASCII-mode, i.e. upper case and lower case characters are treated
differently.
With "Erase field" you can erase its contents.
With "Paste block..." you can choose to paste the block buffer to the end
of a field, to the start of a field, or to fill the field with the block
text and thus completely replace its contents. See also BLOCK FUNCTIONS.
For "Case conversion" see CASE CONVERSION.
With "Remove marks" you can remove all text placed in {}-brackets from
the remarks field.
With "Remove abstracts" you can remove any abstracts that are stored in
the remarks field. The abstracts that are stored in the remarks field by
the conversion programs included in this release of REFBASE will always
start with the text string "---". The "Remove abstracts" option simply
searches for "---" in the remarks field and then removes all text
including and after this text string. If you wish you can enter a
different start string to search for. If you enter an empty string, the
search will be cancelled.
With "Remove issues" you can remove all issue numbers that are placed in
()-brackets at the end of the journal field.
With "Import ... " you can import a field in the currently loaded REFBASE
file. The import file must be a text file in which the number of lines is
equal to the number of records in the REF_BASE file. The import facility
will then place each line of the import file in the selected field of the
REFBASE file, starting with the first record that is visible in "Display
All" mode.
---------------------------------------------------------------------
MODE MENU
Most of the options of the Mode menu have already been discussed in the
other sections of the manual:
"Cursor blink" with this option you can specify whether or
not the cursor should blink during data
entry
"Use wildcards" see SEARCHING/SELECTION
"Recall previous sorts" see SORTING
"Sort/Goto/FindDupl by ASCII" see TEXT MODE
"Partial equation" see SEARCHING/SELECTION
"Use quickfind algorithm" see SEARCHING/SELECTION
"Mouse delay time" with this option you can enter the value
(0-9) of a short time pause to be given
after each mouse function (e.g. mouse
click, change from arrow to bee etc) in the
program. The default value is 5. See also
THE MOUSE.
"Soft scroll pause units:" see DISPLAY
"Use VDI 8*16 font" With this option you can specify whether
the VDI font should be used for displaying
records in the standard 8*16 characters
(instead of the TOS font). This further
increases the scrolling speed if a screen
accelerator is installed that specifically
speeds up the VDI screen output.
In the main menu you can save the MODE configuration by clicking on "Save
MODE preferences ... " or pressing "M". This will create a file
"REFBASE.MOD" in the same (sub)directory that REFBASE was started from.
Whenever REFBASE is loaded, this file will now be automatically read and
the preferences set accordingly.
---------------------------------------------------------------------
THE MOUSE
Programs written in GFA-Basic often have difficulties with the standard
GEM alert boxes: the default button becomes black as soon as you move the
mouse above it, you may need to click twice before the alert box
responds, and sometimes pressing RETURN even selects the non-default
option instead. For this reason REFBASE uses its own alert box, which can
be operated with both the left and right mouse button. In fact, all
mouse-selectable functions in REFBASE, except for the GEM-fileselect box,
respond to the right mouse button as well.
Users of TOS version 1.4 may have installed the button fix accessory
(BUTTNFIX.ACC) to overcome some problems with the mouse-click in GEM-
programs. Unfortunately, using BUTTNFIX.ACC affects the mouse functions
in REFBASE. Examples are sudden pauses or halts in the program (which can
be overcome with an extra click of the mouse), erratic and unwanted
selections in the GEM-fileselect boxes, or difficulties in returning to
the Desktop after quitting. These errors can be prevented reasonably well
by introducing a small mouse delay time. The length of this time pause
can be set in the mode menu. The alternative is not to use BUTTNFIX.ACC
together with REFBASE, or to use the right mouse button for selecting the
various functions instead. In the latter case the mouse errors will not
occur.
---------------------------------------------------------------------
KEYWORDS
You can enter the keywords menu from the main menu by clicking on
"Keywrds" (SHIFT-F3), or pressing "K". You can also enter the keywords
menu during "Display" or during data entry ("Input"/"Amend") by pressing
"K" or CONTROL-K respectively.
From the keywords menu you can load a file with keywords (or if necessary
any other text file) by pressing or clicking on F3. You can scroll
through the file by pressing the CURSOR UP and CURSOR DOWN keys. With
CONTROL-CURSOR UP/DOWN you can scroll by page. CONTROL-CLRHOME and
CONTROL-Z jump to the top and bottom of the file respectively. Each line
is numbered to give an indication of the current position in the keywords
file. The numeric keypad can also be used to scroll through the keywords
(see DISPLAY).
If your keywords are sorted, you can jump through the file by binary
search with the GOTO command (F7). GOTO normally works in alphabetical
mode; if your keywords are sorted by ASCII-code, you must first activate
the ASCII sort option in the mode menu, otherwise GOTO will not work
correctly.
With "Select" (F2) you can search for all keywords that contain a
specified text string. You are allowed to make one combination of search
terms by typing in "AND", "OR", or "NOT". Start the search by pressing
RETURN. If no matching keywords are found, you return to the main list;
otherwise the selected keywords are shown, and you can return to the main
list by pressing ESC. The "Select" option can be set in alphabetical or
ASCII mode by pressing F4. (This is the same switch as SHIFT-F5 in the
main menu). With Find (F5) you can search for the first keyword that
includes a specified text string. The cursor will then move to this
position in the keywords file. You can repeat this search by pressing F6.
With "Erase" the loaded keywords file is erased and the memory used is
available again. Also when a new keywords file is loaded, the preceding
one is erased first. You can quit the keywords menu by pressing ESC or
"Exit" (F8).
While entering or editing records, keywords can automatically be inserted
at the cursor position. To do this, you must press CONTROL-K to enter the
keywords menu, move to the desired keyword, and then press RETURN. Repeat
this sequence if you want to collect more keywords. Each inserted keyword
will be terminated with "; " as a separator. It is advised that you do
this also when you type in the keywords yourself. Only the last keyword
in the keywords field should be free of this separator sequence. In this
way you can use keywords which actually consist of two or more words
separated by spaces, e.g. "polyacrylamide gel electrophoresis". With
KEYWORDS.PRG you can make a sorted list of your exported keywords, which
can be loaded into REFBASE. This program has a special facility for
recognizing keywords that are separated by "; ".
NOTE: keywords containing a double space (" ") will be inserted
incompletely when pressing RETURN, since all text on the right side of
the most right double space will be removed. This enables you to load
keyword files in which each keyword is provided with its word frequency,
separated by a double space. For instance, the keyword "monoclonal
antibody 10" indicates that this occurred 10 times in the extracted
keyword file; if you press RETURN, the "10" will be stripped off since it
is preceded by a double space, and only "monoclonal antibody" will be
inserted into the record. KEYWORDS.PRG offers you the option to make such
keyword files with word frequencies. (Normally, keywords entered in
REFBASE will never contain double spaces, since they are filtered out
after data entry).
The keyword files use space that is shared with the records. Therefore,
if large files are loaded the keywords may be overwritten by the records.
The user is notified when this occurs, and the use of the loaded keywords
is blocked. Loading of keyword files is cancelled if there turns out to
be insufficient space or RAM. Loading of keyword files will never lead to
overwriting of records in memory.
----------------------------------------------------------------------
PRINTING
Clicking on "Print" (SHIFT-F4) gives the print menu. You can leave this
menu again by pressing RETURN, of clicking on EXIT. To activate the
functions in the print menu simply click on them with the mouse.
With the print menu, reports can be generated from either all, selected
of tagged records. These reports can be sent to printers connected to the
parallel port (= printer port) or the serial port (= modem port), written
to disk, or shown on the screen. The layout of the reports is essentially
the same for each of the four "print directions".
You can choose between the following report types:
1. Reference = author + year + title + journal. In addition to this
short format, you can also select a medium format (including the
File and the Subject fields as well), a large format (to which the
keywords and the marks in the remarks field are added) and
finally the full format (which includes the remarks field in
full).
2. Card - all fields are printed including labels, with each field
starting on a new line.
3. Author Table - A table with authors + year + subject + journal +
date + read + file. If you choose a column width of 136 (17
characters per inch) instead of 160 (20 c.p.i) then the date
field is omitted.
4. Title table - Here you can choose for titles only, or for a more
complete table with (very short) authors, year, subject, journal,
and file as well.
5. Journal table - This will only print journal + year + file. If you
choose to print in pages, the table will be printed in two
columns.
Start printing with "Print Report". Stop printing by pressing ESCAPE.
By clicking on "Continuous Print"/"Pages + formfeeds" you can indicate
whether or not you want to print in pages.
If you print in pages, (1) each page can be printed with a header, (2)
between pages the printer is instructed to move to a new sheet of paper
(= form feed), (3) Journal tables are printed in two columns, (4) it is
possible to print only even or uneven pages, thus facilitating double-
sided printing, and (5) you can also choose to pause printing between
pages. Furthermore, if the last reference or card of a page does not fit
completely at the bottom of the paper, it will be printed at the start of
the next page.
With continuous print, no form feeds are given and the records are sent
"straight" to the printer.
Click on "No header"/"Print header" to indicate if you want page headers
to be printed. With continuous print, only one header will be printed.
Click on "Show header..." to enter or modify the page header. Each header
will automatically contain the system date and the page number, and the
user can specify any additional information.
Click on "Print margins/No margins" to indicate if you want a margin to
be printed. This margin will allow sufficient space for a paper
perforator. You can place the margin either on the left side or the right
side, or if desired left on the uneven pages, and right on the even
pages.
With "card print" you can select the label-text that will be printed
before field 4: "DOCU" or "READ".
You can choose between two printer types:
1) The standard 9-pin matrix printer (24-pin will probably also work
properly although this has not been tested)
For this printer type you may choose between 10, 12, 17 and 20
characters per inch (= pica, elite, condensed, and elite
condensed respectively). Furthermore you can choose between 7
and 12 lines per inch; in the latter case the text will be
printed in superscript. Other selectable options are Draft/N.L.Q
and the paper height (11 or 12 inch). To prevent the columns in
the tables from being printed too irregularly, an alert box will
offered you the possibility of unidirectional print as well.
2) The HP-Deskjet printer.
The following fonts are selectable:
(cpi = characters per inch; lpi = lines per inch)
Cour 10 cpi 6 lpi - standard Courier
Pres 12 cpi 6 lpi - standard Prestige
Pres 12 cpi 7 lpi - standard Prestige, reduced line spacing
Pres 16 cpi 8 lpi - small Prestige
Pres 16 cpi 10 lpi - small Prestige, reduced line spacing
Cour 20 cpi 12 lpi - small Courier
For the Prestige fonts you will need to install the Prestige font
cartridge in your printer.
For printing tables, you are obliged to choose a character pitch of 16,
17 or 20 cpi, even if you only want to print to disk, or to the screen.
For printing cards or references, you are free to choose any character
pitch you want.
If you want to export references to a file, it is best to choose 10 cpi
with a right margin. The references will then become about 75 characters
wide, and they will fit nicely in your word processor documents. If you
load reference files in 1ST Word Plus, you must first activate the WP-
mode (if it was inactive) and then replace every space (" ") by a space
(" "). This will convert the references to the Word Plus format, and they
will now respond to the "reformat" key.
When printing references, the journal field is printed in boldface
(matrix printer) or underlined (Deskjet).
If you print to a printer connected to the serial port, you must first
make sure that the "RS232 PORT CONFIGURATION" is properly installed with
the control panel accessory. For the HP-Deskjet the following settings
will work OK: Baud Rate "19200", Parity "None", Duplex "Full", Bits/Char
"8", Strip Bit "Off", Xon/Xoff "On", and Rts/Cts "Off".
----------------------------------------------------------------------
CHAIN SEARCH
You can enter the Chain search menu from the main menu by pressing "C" or
clicking on "Chain search ...". You leave the chain search menu by
pressing "E" or ESC, or by clicking on "Exit chain search ..."
With the chain search multiple files on disk can be searched for records
that match a given selection rule. This search facility only makes sense
if you have more records than can be loaded in the memory of the
computer, and if you have a hard disk.
By clicking on "File" (F1) you enter the chain search file-select menu.
Select the files you want to search as described in the LOADING FILES
section. Accept a selection by clicking on CHAIN. Note that all files to
be searched must be in the same (sub)directory, and that each separate
file must fit in the ST's memory. A warning is given if any of the files
is too large to be loaded. The maximum number of files in the directory
(and thus for the chain search) is 68.
Use F2 to choose a selection rule as described in the SEARCHING/SELECTION
section. With F4 ("Alpha"/"ASCII") you can choose whether or not
distinction between upper and lower case characters must be made during
the search (See TEXT MODE). The selection rule remains active until
replaced by another, and is not affected by any selection rules that are
entered in the main menu.
To start a chain search press or click on F3 ("Search"). Choose "Count
selected records" if you only want to count how many records match the
selection rule; choose "Fetch selected records" if you want to fetch and
see them. During the chain search the selected files are loaded one after
the other, and the number of records matching the selection rule (hits)
are displayed for each file on the screen. After the chain search the
total number of hits is shown in an alert box, and if you had chosen to
fetch the selected records, they can be displayed.
If records are fetched during the chain search, they are first stored in
a RAM-buffer, the size of which is indicated in the chain search menu. If
this buffer becomes full, REFBASE will start writing the search results
to a separate file on disk. The maximum size of this file is indicated by
the "Fetch limits" in the chain search menu, and depends on the amount of
memory in your computer. As soon as this maximum file size is reached,
the chain search will be terminated. You can modify the drive path for
saving the chain search results by clicking on "Write drive", and alter
the name of the file by clicking on "Save overflow as...". After the
chain search, the selected records are displayed either after reading
from the RAM-buffer, or after loading from disk.
If during a chain search a loaded file turns out to have too many
records, and thus does not fit in memory, the chain search will be
stopped.
Pressing/clicking on F5 ("Hitlist") will show the selection rule of the
latest run, and the number of hits in each of the searched files. You can
scroll the list with the cursor keys. Exit with ESC or a mouse-click.
---------------------------------------------------------------------
FILE FORMAT / DATA EXCHANGE
REFBASE stores its records as plain text files, with each field occupying
a single line, and each record consisting of 10 lines. REFBASE files can
therefore be loaded in every text editor or word processor. A problem may
arise though with the block fields. If their length exceeds the maximum
line length of the text editor (or word processor), then they will be cut
off, and the remainder of the field is placed in the next line. For
TEMPUS the maximum line length is 255 characters, for 1st Word Plus it is
150. Loading REFBASE files in these programs may therefore damage the
file structure, and if you save them again they may become useless for
REFBASE. If you want to import REFBASE files in other databases, these
programs must be able to read lines that are longer than 255 characters.
For those who want to write conversion programs in GFA-Basic V3, be sure
to load the REFBASE files with the "RECALL" function. If you use the
"LINE INPUT" function, long fields will as yet be truncated at 255
characters, and their remainder passed on to the next line that is to be
read from disk.
--------------------------------------------------------------------
SUPPORTING PROGRAMS
With this release of REFBASE four supporting programs are provided as
well:
1) KEYWORDS.PRG - With this program you can create a sorted list of
keywords from your references, which you can subsequently load in
the keywords menu of REFBASE. First export the contents of the
keywords fields to a file on disk. Then read this file with
KEYWORDS.PRG to make the list. If you have separated your keywords
with a semicolon + space ("; ") you must indicate this in the menu
of KEYWORDS.PRG before you start the extraction. Otherwise, by
default, the space is will be interpreted as the keyword separator.
2) CC_HELP.PRG - with this program you can convert export files from
the Current Contents on Diskette (CCoD) into files that can be read
by REFBASE. To do this, you must export your CCoD search results in
the "RAP comma-delimited" format. The fields of the REFBASE files
will then be filled with the following information:
SUBJECT = Discipline (abbreviated according to internal tables in
the conversion program; if a discipline in the RAP file does not
match any of the preset values it will be reported in an alert box,
and the discipline will be marked with an asterisk).
READ/DOCUMENT = Document/Language in abbreviated form
FILE = The first few characters of the name of the export file (see
CC_HELP.PRG)
JOURNAL = journal + volume + pages + issue. (a standard format is
used as explained in CC_HELP.PRG)
DATE = The date of the journal issue (if given in the export file)
REMARKS = The address of the author to whom reprint requests should
be sent. (This address is always terminated with a "|", to
facilitate its recognition by any "reprint request programs" that
the user may wish to write).
The other fields need no explanation
CC_HELP.PRG also offers the option to load and print your CCOD
search profiles. If you intend to routinely collect the full
contents of your favourite journals, it is advised to create a
separate profile with only "Journal" fields. After exporting records
from CCoD with this profile, and converting them to a REFBASE file,
you can use the issue extraction facility of CC_HELP.PRG to collect
the issue numbers in this REF-file.
Issue files are recognized by REFBASE from their extension ".ISS".
When they are loaded in REFBASE, the "Expand" option of the table
view is inactive, and only a single screen table with an adjusted
selection of fields is shown. Also sorting on the journal field will
go differently. For instance issue "(9)" will now be sorted before
issue "(10)", and "(JAN)" will be sorted before "(FEB)".
3) FULMERGE.PRG - With this program you can merge abstracts and
additional book-information from the Current Contents on Diskette
into your REFBASE file.
From the "CCoD with abstracts" you can export your records in the
so-called "Full + abstracts" format, which not only includes
abstracts but also the complete information (titles, editors,
publishers) of books from the "Current Book Contents". Unfortunately
this format is not suitable for direct conversion to the REFBASE
format, because all journal names are given in full. This would make
the journal fields too long.
The "RAP comma-delimited" export format however does have
abbreviated journal names, but lacks both the abstracts and the
complete book-information. The solution to this problem is to export
your records twice from the CCoD: once in the "Full + abstracts"
(extension ".FUL") and once in the "RAP comma-delimited" format
(extension ".RAP"). First you convert the ".RAP" file to a REFBASE
file with CC_HELP.PRG. Then you use FULMERGE.PRG to merge the
essential data (abstract + book information) from the ".FUL" file
into the REFBASE file. This will work o.k. as long as you do not
change the sort order of the records in REFBASE file before the
merge operation. You now have records that combine abbreviated
journal names with abstracts and full book information.
The abstracts will be placed in the remarks field, immediately after
the reprint address, and they will be preceded by the separator
string "---". You can later on always automatically remove any
unwanted abstracts with the help of this separator string (see
FIELDS MENU). The complete book information is placed in the title
field, also preceded by a "---" separator string.
4) REFSPLIT.PRG - Use this program to recover REFBASE files that have
accidentally become too large to load in the computer's memory (e.g.
due to a save + append operation). REFSPLIT.PRG splits the file in
two or more smaller parts, which can then be loaded separately
again.
----------------------------------------------------------------------
LIMITS OF REFBASE
The maximum record size is 5000 characters. With larger records errors
may occur during saving. When references from the Current Contents on
Diskette are converted to REF_BASE format with CC_HELP.PRG or
FULMERGE.PRG, their size will be tested, and they will be shortened if
they have more than 4500 characters. (REFBASE itself does not yet check
whether records exceed the maximum size).
The maximum number of records that REFBASE can display is 32750. This
limit is expected to be reached only on Atari-ST compatible computers
with more than 20 Mb of RAM. In practice the maximum number of records
that will be loaded (= "record space") ranges from about 600 for a 1 Mb
ST to 5800 for a 4 Mb ST.
The maximum number of files and folders that can be displayed in the load
menu is 68. This is also the maximum number of files for a chain-search.
The maximum number of keywords that can be loaded is 10 times the record
space; the maximum number of selected keywords that can be displayed as a
subset is equal to the record space.
The sorting algorithm has been programmed recursively. I have no idea how
many levels of recursion are allowed on the Atari ST, nor do I know how
many levels or recursion normally occur during a sort. It may be possible
that this recursion poses a limit on the maximum number of records that
can be sorted. However, I have experienced no difficulties in sorting
5000 records on a Mega ST4.
REFBASE was written in BASIC. This implies that from time to time the
user may expect short pauses or response delays in the program, which are
caused by an automatic memory cleanup operation called the "string
garbage collection". This phenomenon passes almost unnoticed with 1 or 2
Mb machines, but on 4 Mb machines with only a few hundred kilobytes of
memory space left it may take one or two seconds before such a "freeze"
of the program stops and action is resumed.
---------------------------------------------------------------------
End of text