Category : Databases and related files
Archive   : RSX24.ZIP
Filename : RSX.DOC

REFERENCE
MANAGEMENT
SYSTEM


Version 2.4h



A Personal System for Managing
Entry, Search, and Report of
Bibliographic Data



Reference Management System provides full file
compatibility with Ashton Tate's dBase III, and
Wordtech's dBXL. It provides a self explanatory,
menu driven system for handling many important
data entry, viewing, editing, sorting, searching,
and reporting operations.



Information in this document is subject to change without notice and does
not represent a commitment of any kind from the author or distributor.

The software described in this document is distributed at minimal charge
for evaluation only. Users of this software are to license this software
according to the instructions contained in this document.

Copies of this software may be made only in accordance with the
instructions given in this document.


(C) Copyright Paul Douglas Goodell,
1987,1988,1989,1990,1991
All rights reserved.


Some portions of this software are copyright by WordTech Systems Inc.
and are used in accordance with their distribution requirements.

Some code concepts copyright by Sybex,Inc. are used with permission.

dBASE, dBASE II, dBASE III, dBASE III+ are registered trademarks of
Ashton-Tate Corp.

dBXL and QuickSilver are registered trademarks of WordTech Systems,Inc.


MANUAL OUTLINE


LICENSING AND DUPLICATION
Duplication and General Information
Licensing Information

INTRODUCTION & INSTALLATION
Background
General Description
System Requirements
Program Files
Definitions
Installation
Set-Up

USING REFERENCE MANAGEMENT SYSTEM
Introduction
Full Screen Record Display
Opening a Database & Index
Record Management
Adding & Editing Records
Rapid Record Scan
Duplicate & Deletion Options
Search Techniques
View/Search
Single Item Search
Multi-Item Search
Document Generation
Reports
Topical Lists
Utilities
Miscellaneous Notes

APPENDIX
Database File Structure
Special Thanks
Yet to Come



DUPLICATION AND GENERAL INFORMATION

Reference Management System is distributed in both 'shareware' and fully
licensed forms. It may be freely copied and distributed for evaluation as
long as five basic rules are followed:

1. The program and this documentation may not be modified
in any way.

2. A copy of this documentation is provided with the program.

3. No charge, other than that for "media and handling" is
made (not to exceed $8.00/disk).

4. Users of the program purchase a license for its use.

5. Registration code numbers issued to licensees are not copied
and distributed.


One license is required per usable copy of the software. Archive and
backup copies do not count.

Note: if you have purchased a disk containing this software from a public
domain/shareware distributor, you have paid the media and handling charges
permitted above. You have not purchased a license. Purchasing a license
is required for continued use.

Extensive effort has been made to make this software easy to use and error
free. However, there are no warranties, expressed or implied, including
warranties of fitness for a particular purpose associated with this software.
Any such warranty is expressly disclaimed, as is any liability for direct or
consequential damages arising from use of this product. This software is
offered on an 'AS IS' basis, and users accept full responsibility for its
effective and intelligent use.


LICENSING INFORMATION

Reference Management System must be licensed to be used other than for
initial evaluation. The software is fully funtional as provided and can be
evaluated completely. License holders will receive information as
available regarding problems, fixes, updates, and application notes for the
software.

An opening screen will identify an unlicensed copy of the software, as will
a notice in place of a users name appearing in various menus. These are
reminders in the event you continue to use the software; they do not
interfere with its operation. Some distributions may have an internal date
stamp that can affect operation at some point. These reminders are removed
upon entry of the license number.

License Purchase

Rates for payment with order:

$45. Single user license with disketts and printed documentation.

$10. Software only on diskette, unlicensed, for trial.

Please specify 5 1/4" or 3 1/2" disk size.

Personal and business checks are accepted. List items desired
and include name and address.

Rates for Purchase Orders requiring billing, or for purchase via third
party vendors:

$60. Includes single user licensed diskettes
(5 1/4" or 3 1/2") and printed documentation.
Site license available on request.

Purchase orders are subject to acceptance. Orders obligating
your company's terms and conditions will generally be rejected.

Enclose Purchase order stating:

Number of copies Also include address and phone of
Shipping address individual to be contacted should any
Billing address question arise about the order.
Diskette size

Mail orders to: P. D. Goodell
234 Mulberry Place
Ridgewood, NJ 07450


WHY REFERENCE MANAGEMENT SYSTEM

A little personal background for those who might be curious about the
origin of this software. The author is a scientist (metallurgist)
frequently engaged in literature searches and preparing articles for
publication. Having used the PC for many other tasks, I had been
frustrated by its handling of bibliographic information (circa 1985).
I wanted a simple set of tools, but also I wanted the literature handling
done in a certain way. If you want it done your way, as they say . . .

Bibliographic software for PC's was extremely limited so I turned to the
general purpose database managers (DBM's). The commercial flat-file
databases were inadequate for certain input types and report flexibility.
The relational database products offered were not necessarily needed for
this application, but some included command languages that provided an
interesting way to deal with varying requirements.

There were several good DBM products at the time, and I focused on Ashton-
Tate's dBASE III for several reasons. One was the availability of third
party support which was very much needed. Another was the growing
portability of the dBASE file structure -- I certainly did not want a lot
of effort locked into a fixed product. This has proved valuable and there
are many DBM's and utility programs that now read and write both dBase III
data and index file types. This package was written using dBase III
and/or dBXL.

The third reason for dBase-type selection came somewhat later as compilers
for the dBase language improved and became attractive. Compilers offered
speed of program execution. They also provided the means to stand-alone
application development. WordTech Systems' QuickSilver was selected by
reason of complete data file compatibility, unrestricted distribution, and
extensions to the dBase language.

But, back to the starting objective, the bibliography work. I had not
wanted to fuss continously with the mechanics of dBase type operations. So, I
decided to write a few small programs to automate common operations. Well,
those few small programs took on a life of their own, especially as extension
to other users was requested. The activity developed into a full scale
independent project. Utilities were added, search operations were extended,
report capabilities were expanded, checking and error trapping were introduced.
It ended up as the complete product you see here.

While the mechanics of the program were being worked out and debugged, the
display interface was esentially left alone. It is a straight menu selection
type interface with no bells and whistles -- no pull-down menus, moving light
bars, etc. These might be introduced in future versions. For now we have a
solid base that has been built and evaluated by extensive use. It has worked
well for me and others. I hope it will also help in your work. I will continue
to attempt improvements, so I would appreciate any comments you may offer.

P. D. Goodell


GENERAL DESCRIPTION

Reference Management System provides a self explanatory, menu driven system
for bibliographic reference handling: data entry, viewing, editing, searching,
screening, tabulation and reporting. It provides data file compatibility with
Ashton Tate's dBase III and Wordtech's dBXL so that data can be easily utilized
elsewhere. It also provides greater speed for many operations than is possible
in non-compiled applications.

Its purpose is to free the user from the need to fuss with set-up and
programing requirements of data base programs. However, since all files
are compatible with dBase III, dBXL, etc., it leaves all the power and
functions of these available for use, possibly for actions beyond the scope
of this program.

Reference material in files is managed as records which contain data fields.
There are fields for author(s), title, source (publication, volume, pages,
year), key words, sponsor, comments, special codes, and abstract. All except
abstracts can be searched upon by several techniques to find records if
particular interest. Records can be indexed or sorted by various key fields,
and checked for duplicates. Individual records and subsets of the database
can be extracted to create new databases or reports; extraction can be
specified by any user-designated order or search-match criteria.

Lists can be generated to show frequency of occurrence of certain data (key
words, author, etc). Reports can be stored to disk-file for incorporation
into word processor texts, or printed in hard copy form. Report formats (disk
or printed) can be varied to accommodate different publication requirements
and can be either line list type or index card type.

In addition to the basic data management, there are several utility
functions. These include file manipulations such as create, erase, append,
and check status. Limited printer manipulation is also provided. There
are facilities to import and export data of different format types.

Reference Management System is organized into five basic menu screens each
of which cover four or five related operations. These are described later
in this documentation. However, the best way to get a feel for the program
capabilities is to try it out using the default settings and the demo
database which is included.


SYSTEM REQUIREMENTS

Reference Manager System will run on fully IBM compatible systems. The
program can be run from floppies, but a fixed disk is highly recommended.
For a fixed disk, least 500 kbytes should be available for the program plus
whatever will be needed for your data files.

The config.sys boot-up file must contain the lines files=20 or more and
buffers=22 or more; you may also need to include driver=ansi.sys. If you
are unfamilar with these items, please consult your DOS manual.

The recommended memory level is 512 kbytes, although as little as 490 can
suffice (for example if you are running in a window). More memory can
improve speed of operation. The memory use is allocated as follows (for
those interested):

Reference program requires 225 kb
Text editor for abstracts requires 120 kb
Program variables and data base require 155 kb
Total (approximately) 500 kb


dBase III, dBXL, or similar database management software is not required.
You may, however, wish to use such products to do operations that are not
offered in Reference Management System. It is one of the reasons that the
data file compatability was considered important. Furthermore, though not
necessary, there are some dBase type commands that can be utilized within
Reference Management System if you are familar with the language syntax.

You may also wish to bring reports from Reference Management System into a
full featured word processor for final document preparation. This will be
necessary for introduction of bold, underline, and supscript attributes.
Most other formating features can be handled directly.


PROGRAM FILES

Reference Management System uses and generates several types of files which
are listed here for general reference.

Required Files:

Base program file RSX.EXE

Program overlay files (5) RSC.OVL RSU.OVL
RSM.OVL RSS.OVL
RSR.OVL

Memory data files MEM?.MEM COL?.MEM

Data & list structure files MASTER?.DBI

Error messages ERROR.TXT

Text Editor [(C)-Wordtech Sys.] WEDIT.EXE


Generated and Optional Files:

Data record files name.DBF name.DBT

Data file index name.NDX

Frequency list files name.D0#
name.N0# name.MEM

Demonstration files DEMO.DBF DEMO.DBT
DEMO.NDX DEMO.MEM
DEMO.D0# DEMO.N0#

Manuals RSX.DOC WEDIT.DOC

Transfer Formats (if exist) *.FRM

Installation INSTALL.BAT *.00?
CHOOSE.COM**

**CHOOSE.COM Copyright (C) 1985, Compute! Publications,Inc.,P.O.Box 5406,
324 West Wendover Ave, Greensboro, NC 27403. All rights reserved.
Reprinted by permission.


DEFINITIONS

A few definitions can avoid confusion and may help in understanding how
Reference Management System operates. Most of the terms have straightforward
meanings, but they may be somewhat different than in other systems.

DATA BASE is a specific collection of records. It has a filename (no more
than 8 characters) and an extension ".DBF", which is often not
shown. Database files are saved in the data directory.
Databases must be "open" to be used.

RECORD is an item in the data base consisting of field entries.

FIELDS are the specific items in the record, author, title, journal,
year, key words, etc. Data fields are fixed in length.

OPEN FILE Most functions require an open data base file. Files can
be opened in most main function menu screens. Once open, a file
is available for subsequent functions. It is closed only if
changed or exit from the main menu.

INDEX is a particular order of records of a data base file, arranged
alphabetically or numerically on a field. They can be based on
more than one field. Index files are named and saved to disk
(with ".NDX" extensions). An index used with an open database
file, and will be constantly updated as the database is altered.

RECORD NUMBER is assigned to each record. This number can "float" if
the data base is modified, as for example by deleting records.
Therefore, records can also be given a fixed ident_no.

RECORD/FILE SIZE. The reference record takes about 760 bytes and
unnecessary additions or file creations can use up disk capacity
quickly. As capacity becomes limited, checking and clean-up via
Screens 2 and 3 can be very helpful. (Note, index files don't
take much room.)

MEMO FIELD is a special type of data field that is of variable size.
As such it allows for extended data entries without requiring
excessive disk space. The field used for ABSTRACTS in Reference
Management System is a memo field. Memo fields are saved under
the same file name as the database file, but with a ".DBT"
extension. The memo field is entered by `F4'. This activates an
editor or viewer, depending on the operating mode. Searching is
not possible in an abstract field (yet) and printing is slow.

FILE LOCATION. Files and programs may be on different disk drives
and/or paths. For a floppy disk type system, the default drives
for programs and data are A: and B:, respectively. For a fixed
drive system, C: is the default is drive and the sub-directory
paths for program and data are C:\rsx\ and C:\rsx\lists\,
respectively. These can be changed, but only through the main
menu, option 7 or 8.


INSTALLATION AND SET-UP
BACKUP

The first thing you should do is to make a backup copy of the Reference
Management System distribution diskettes. Use the DOS diskcopy command to
include subdirectories from the program disks.

INSTALLATION

Installation of Reference Management System is not related to any copy
protection system, but simply sets up directories (on your hard disk) into
which the program files are copied. It also helps you to select an
appropriate set of colors for your type of monitor. All initial parameters
can be changed from within the program.

These operations can be done automatically by running the installation
program from drive A. Put RSX Distribution Disk 1 in drive A and type A:
and then type INSTALL ; answer the questions as they appear.
The options relate to selection of operation from a fixed disk or floppy disks,
and to selection of monitor types. Direct installation to a hard drive with
default directories and monitor=color can be done by entering INSTALL #.

Fixed Disk Operation:

For fixed disk operation the two subdirectories created are: \RSX and
\RSX\LISTS. The files listed as 'require' in the tabulation of
Program Files are copied to the \RSX directory. The demonstration
files are copied to the \RSX\LISTS directory, and this is where it is
recommended that the data files be stored.

Floppy Disk Operation:

Reference Management System can be run from a floppy base system, but
the speed of some operations will be reduced. The disk set up will
depend of the type of floppy disks that can be used.

High Density Floppies. If your system can use high density floppies
(1.2 Mb or 760 kb), you may want to set up a data subdirectory
\LISTS and structure the file distribution as described above for
fixed disk operation. You can also operate with all the files in the
root directory; simply identify the appropriate file locations in the
set-up operation. Data files can be on the unused portion of the
working disk or on a separate disk/drive if available.

Low Density Floppies. If you wish to use low density floppies (360
kb), you will need to prepare two disks. Disk #1 will be a start-up
disk and should contain the following files (* = all such files):

RSX.EXE, RSC.OVL, *.MEM, ERROR.TXT

Disk #2 will be the working disk which may replace disk #1 after
starting, and should contain the following files:

*.OVL, ERROR.TXT, MEMTRANS.MEM,
*.DBI, WEDIT.EXE

Data files including the DEMO.* files can be on working disk #2, but
it is better to locate them on a separate data disk used in drive B:
if one is available.

SETUP

Most likely the program will need no further setup and can be started
immediately. Simply switch to the drive and directory containing the
program files and type:
RSX

This will start the program. A license notification will appear as the
first screen until your registration number is entered. Instructions for
this entry are included with the registration. The main menu will follow.
The menus will be discussed later; for now, note that items # 7 and # 8
involve changing the data file location, or any/all of the set-up parameters.

The default parameters established during installation will be usable in
most cases, but they are also a likely source of problems. The data file
location can be changed from within the main menu; simply select item #7
and then fill in the path name. Control of other parameters is by
selection of item #8, which introduces parameter set-up screens.

In the parameter set-up screen, the first set of parameters includes
specification of the drive and path for the program files and the data
files. The second set of parameters covers the screen color characteristics.
You may choose MONO, GRAY, or COLOR; for COLOR, you can accept the existing
color pattern or indicate that you want to change it. Note that the pattern
for COLOR can also be used for MONO & GRAY screen, they are not exclusive.
Color settings can be very effective with a monochrome screen that represents
colors as shades of gray.

Entering the combination type=COLOR and change=Y will bring up a color
selection screen. (You should look at some of the screens before you make
changes.) The various levels of headings can be set for different emphasis.
Most useful among these settings is that of the input requested color;
uniqueness in this setting will help highlight input requests throughout the
program and thereby increase the obviousness and smoothness of program flow.
For monochrome screens, reverse video is recommended (N/W).


USING REFERENCE MANAGEMENT SYSTEM

INTRODUCTION

The reference system is organized as five operating menus. Each Menu and
Sub-Menu function is accessed by entering its number, 1-5, in a highlighted
selection box and pressing enter (Carriage Return). An overview of the
organization shows the menu relations:

System Outline, from MAIN MENU:

1. 2. 3. 4. 5.
INFO UTILITIES MANAGEMENT SEARCH REPORTS
Terms Information Open File Information Information
Outline File Status Scan Records Open File Open File
Notes File Operation Add Records Scanning Topics
File I/O Edit Records SingleItem Lists/Cards
Printer Set-Up Check Dups. Multi Item Word Proc.
Return Return Return Return

Menu screens are numbered for easy reference as for example, screen 5.4 is
data report, output to printer (lists or cards). Each menu includes an
information section which presents purpose and operation of the menu items.
This is option #1 in several menus, and is included within several items of
the management menu. Each menu includes an option to return to the main
menu (option #6).

There is some redundancy among the menus to minimize switching. For
example, the Utility menu contains an option to open, change or create a
data base file, and the three main operation menus each include similar
options. Printer set-up is in Utility (2.5), but some printer options are
also adjustable in the report output section (5.4). Scanning of records is
available in some form under both the data management (3.2) and search
(4.3) operations.

Selecting options and entering responses is indicated by a command line or
question followed by the cursor, usually in a highlighted entry box. The
entry boxes may switch color to indicate when and where an entry is required.
Most choices presented have a default value. These are shown by characters
in upper case (Y,n), or <2> brackets.

The entire system can be reviewed by scanning the menus in default modes;
this is recommended for a fast introduction to the system.


FULL SCREEN RECORD DISPLAY

Data entry and editing are done in a full screen display which presents all
the data of a record as well as some status and control information. It is
central to the system operation and will be reviewed before considering the
specifics of its access.

The status line is the top line of the display. At the left side it shows
the name of the open database file and any index file that may be in use.
The right side shows the currently permitted operation which may be one of
three types: add records, edit data, or view data. The center of the status
line shows *ok* or *del* which reflects a normal record or one marked for
deletion. The state of the toggle between typeover and type-insert mode is
also shown included in the status.

Control information for the cursor and function keys is shown as a block
below the status line. This identifies many important ways to move within
and between records. In addition, the home and end keys will move the
cursor to the start and end of a data line. Note that there are replicate
actions, especially between the function keys and the key pad; use
whichever you find most convenient.

The main part of the display is the data presentation which shows the
complete set of fields for the record. The shaded areas indicate the data
fields. Entries can be as long as the highlighted box, but no longer; they
must be left justified or they won't be found in searches. Use upper and
lower case alpha or numeric characters as desired.

Some data fields are split between two or three lines. This allows room for
the particular type of data, but the entry is not automatically word wrapped
between the lines. Therefore, you may need to do some manual adjustment.

The record number pointer and total records in the file are indicated for
reference and are not entered. The Ident_No., Index_No., and Code fields
are provided for general purpose identification and tracking based on user
defined schemes. It is handy to use one of these fields (Ident_No.) as a
record numbering tool to supplement the internal record number. The internal
record number can float as records are deleted and inserted; a fixed tracking
number can facilitate record extraction for list generation. (A unique Ident_No
is required for some operations.)

A separation is made between first authors and additional authors. The
first author's first and last names are separate to facilitate preparation
of reference lists that differ as to which name listing comes first. This
dual option only applies to the first author listing. Other authors are
included on the next display line and must include appropriate conjunctions
and punctuation within line for multiple authors, but not at the beginning
and end of the list.

Reference source, volume, year, pages, and title are all obvious. The
source and title fields have been made long to accommodate extended
entries. If they are not adequate in a particular case, you might use the
comment field for additional data.

Keywords and sponsors can include multiple entries if they are separated
with comma and space. This will permit searching and listing based on a
series on entries. Key words are one of the most important aspects of
record entries, and are often the least well structured. It can be well
worth the time to construct and maintain a consistent set of key words
relevant to your application. As an aid to this, it is possible to list
key words and their use frequency from the report section of Reference
Management System (see topical lists).

Comments can include up to three lines of whatever notes, annotations, or
comments you care to enter. This area is included as a supplement to the
abstract area, the text of which is not presented on the main display screen.

The abstract field is special type of entry which is included as a memo and
has a different type of behavior. As explained under definitions, a memo
field accommodates extended and variable length data entries. This can be
an abstract or any other entry you wish to make. It cannot be searched
within (this is planned for a future version).

The memo field can be accessed by pressing the function key, F4. It will
be opened for viewing or editing depending on the display mode. Viewing is
fast, but (currently) shows only the last 21 lines (which is usually enough).
Editing can start after a pause during which a full screen text editor is
activated. The editor included here is the Wordtech WEDIT. It has its
limitations, but will get the job done. A documentation file wedit.doc is
included with Reference Management System that gives full details for this
editor. Exit the editor by Esc and select the save option.


OPENING A DATABASE & INDEX

The primary unit of data storage is the database file. This is subdivided
into the individual records with the data fields. A database is automatically
opened when it is created. An existing database file can be opened for use
from within the Utility Menu, option 2.3.1. Database files may also be opened
from the Management, Search, or Report Menus, option #1, in which case a prior
existing index can also be activated. Changing the open database closes any
database and index files previously opened.

Creating a new database file can be done in the Utility Menu under the option
File Generation/Transfer, screen 2.4.1. Alternatively, files can be created
under the Open or Change option of the Management Menu, screen 3.1. In either
case, select the CREATE option and enter a file name when requested. There will
be a check of existing files for the name assigned, and if none exists, it will
be created. The file is then ready to use, but contains no records until data
entry is started in the Management Menu, screen 3.3.

When a database file is created by Reference Management System, the structure
is copied from a master file. The structure has been preset to balance the
needs of bibliographic data (field size) and the size of records and files.
The structure can be viewed in Utility Menu, File Survey, screen 2.2.2.

While creating a database file it can be useful to consider index files.
There may be a frequently used index, e.g. an order on the last name of the
first author. Index files can be generated at any time or at any point in
the data entry process, including prior to data entry. If an index is
active during record additions or editing of a database, it will be
automatically updated; if it is not active, it will have to be regenerated.
For large databases this can take time which can be avoided if the index
started prior to data entry. An index is generated only in the Utility
Menu, File Organization, screen 2.3. A database file may have several
indices, each built on different fields or combinations thereof, however,
only one index can be active at any one time. Note that when viewing an
indexed file, the order of presentation is that of the index, not that of
the entry chronology; this may or may not be desired as you work with the
database.


RECORD MANAGEMENT

Adding records to a database and editing records are treated in the Record
Management Menu which is option #3 of the Main Menu. The management operations
also include rapid scanning of records in the file, checking for duplicate
records, and deleting unwanted records.

A database must be open to use the functions of this menu; the lines at the
top of the menu will indicate the database status. Opening or changing a
database can be done with option #1. You can also elect to use the database
in an indexed mode and enter the index name. Activate the index following
entry of the database name, or by entering '0' in the menu choice box. Adding
and editing records uses the full screen record display discussed earlier.

ADDING & EDITING RECORDS

Records added to a database are appended to the end of the database (whether
or not an index is active). Full access to all of the fields of a record is
available through the full screen display. Selecting the Add Records option,
#3, of the Management Menu, opens short dialogue box. This shows number of
records currently in the file, provides option for information on data format,
provides option for repeated field entries, lets you start or return to menu.

Once in the add record entry mode, fill in any of the entry positions (fields)
for which you have data, leave the others blank. The only exception to this is
the first entry, ident_no, which must be filled in and must be a unique entry,
i.e., the only one of its kind. It could be the same as the record number
unless deletions have been made. Re-numbering of ident_no can be done auto-
matically in Utility Menu 2.3.5. Also, entries must be left justified or they
will not be found in searching operations.

The editing mode of Reference Management System is accessed by selection of
the Edit Records option, #4, of the Management Menu. This option opens
with a short dialogue box in which four choices are presented:

1. Start with page of data format information?(y,N)____
2. Edit records or view only (N=view,Y=edit)? (n,N)____
3. Number of records in file=####, start at number ____
4. Stop and return to the Management Menu? (y,N) ____

Thus you can get instructions, view only or edit, select point of entry, or
have second thoughts and return to the menu.

When adding records you can back up to check prior records if needed, but you
cannot edit prior records in the add records mode; you can edit records just
added but not yet saved. When editing records you can move anywhere in the
file, but you cannot add new records. The status indicator in the upper right
will show the operation currently possible. You should save the new records
and edits with some regularity in order to avoid losses in the event of system
failure.


In the full screen editing mode you can move forward and backward by record
(record number or index position) by using the PgDn and PgUp keys or F6 and
F5. To move by several records or to find a particular record, the ^Home
or F7 keys have an especially valuable role. These keys open a dialogue box
from which any record can be selected. The selection criteria can be record
number (which is the fastest), or an alternate selection can be made by
entering a period. This allows selection based on Ident_No. or Last Name.
This can provide rapid record access for viewing and editing.

The ^PgUp or F9 keys MARK a record for deletion; this changes its status
designation, but does not physically delete the record until the
delete/pack operation is preformed in Check Duplicate option, #5, of the
Management Menu.


RAPID RECORD SCAN

Option 3.2 and provides for rapid scan of records in the data base file.
It is intended as a quick operation if needed while in the Record Management
Menu. It contrasts with the edit/view option which uses the full screen
display for a single record, showing all the fields. The rapid scan includes
four or five records per display screen, but shows only a limited number of
fields. The fields used in the rapid scan are: record number (on first line),
followed by ident-number, authors, publication date, and title of article.
It is also possible to qualify* or filter the scan to limit the display. After
every four screens, you have the option to quit the scan and return to the
Management Menu. There is no printout option at this point, but output is
treated elsewhere.

-----------------
*The qualifier entered may be up to forty characters and must include the

full syntax for a dBase "Locate For/Do While" procedure. This option is
included here as a convenience for the advanced user; the same type operation
is accomplished and more fully supported in the search and report menus.
------------------


DUPLICATE & DELETION OPTIONS

Selection 5 of the Record Management Menu facilitates treatment of
duplicate records and records marked for deletion within database files.
Four operations are used which can be selected independently:

1. Identify & mark duplicate records for elimination.
2. Display records marked to be eliminated.
3. Transfer marked records to separate file.
4. Eliminate the marked records.

A fifth option is to return to the Record Management Menu.

The first option is to check the database file for duplicate records.
Checking is based on comparison of entries in one or more fields. The
procedure is to identify duplicates and mark those that should be deleted -
- they are only marked at this point, not actually deleted.

The criteria for checking duplication is based on a combination of 3 fields:
first author's last name, year published, and first page of the reference.
The default check criteria appears as: lst_name+year+page_frst. This can be
accepted as is, or changed.

The default duplication criteria was selected as having a high probability
of identifying only true duplicates. If you change it, be sure that you
have the correct field names because there is no error trapping here. The
field names can be displayed under file structure in the Utility Survey
Menu (2.2.4) and are listed in the Appendix.

This operation takes some time, so be prepared to wait. During the operation
a temporary index is generated and records are compared three at a time.
Matches are recorded and presented to you to mark for deletion. If there are
four duplicates of the same item, one will not be picked up unless you run two
passes. Also, limitations on the number of memory variables limits to thirty
the total number of duplicates that can be isolated in one pass. You will be
prompted if this limit is reached; you must delete some and restart the check
to identify more.

On completion of the search, the number of possible duplicate sets is
counted. Possible duplicates are then displayed in sets so that you can
make a selection of those which should be marked for deletion.

The second option of the duplicate/deletion menu provides for a display of
all records marked for deletion. The operation presents a short form of
the data from records marked for deletion and allows a confirmation or
change of the mark status.

If you have just used option one, you may wish to skip option two.
However, it does provide a good control step before records are actually
deleted. Note that duplicate checking is not the only place where records
may have been marked for deletion; marks may also have been entered during
editing, and this will be your last review.

The third option is a further safety device. It allows records marked for
deletion to be transferred to another database file, and a hard copy can be
printed in the process. By default, the transfer file is named old_rec.dbf,
but you can change the name as desired. The old-record type file can be
created fresh or appended to an existing file. With this option you can
collect all files that are deleted from the main database.

The fourth option is the actual deletion of marked records from the
database file. There is one final opportunity to abort this operation
before deletion starts. As records are deleted, the file is re-structured
(packed) to close up the holes. The record numbers are shifted accordingly.
This is one of the reasons that the record numbers float and warrant use of
the ident_no if fixed identification numbers are required.


SEARCH TECHNIQUES

The Search Menu provides three search techniques: view/search, unique item
listing, and generalized multi-field search. The first is appropriate to
general viewing of single records, the second and third are appropriate to
complex searches, culling, or ordering of specific groups of records according
to user selected criteria. The Search Menu is selection #4 from the Main Menu.

View/Search

The view/search technique is option 3 of the Search Menu. It is a general
query approach that facilitates movement throughout the database by single
records, in a full screen display mode. It is similar to the view/edit
mode of the Record Management Menu and offers the same type of action:
forward and backward by individual records, by input or indexed order, or
specific selection by record number, ident_no or last_name.

Unique Item Listing

A particular type of Single Field Search technique is provided as option 4
of the Search Menu. Single field searching can also be done in the multi-
field option. What is special about option 4 is that it provides rapid
location and ordering of records based on a unique item. It is most useful
with ident_no or index_no fields which are unique to a single record. Such
items can be entered individually or listed in groups in any desired order.
The database is searched and the records presented in the manner requested.
Only a limited display of the record is presented on the screen, but the
complete record can be copied to a new file. This can be used for full
presentation or transfer to a word processor (see Document Generation).

The main function of this option is to prepare reference lists for use in
text documents. It will pull out a subset of records from a database file,
and list them in an order corresponding to that specified in the text. For
example, your text cites four references which are numbered consecutively
as they occur in the document. They correspond to records having
identification numbers 7, 453, 27, and 87 in your database file. Entering
the character string 7,453,25,87 will create a file of just those records,
properly ordered. This file can be transferred to your text with consecutive
numbering.

The Single Field Search starts with a dialogue screen to define the process
including the field name for the search operation. In the final preparation
there is an opportunity to specify the name of the search-field index file if
it already exists. If it does not exist (or you can't remember its name), it
will be generated, but this may take time if the database is large.

Next follows a request for the specific item(s):

What item is sought in the specified field? Item = _______

Enter a single item or a list of items separated by commas. After the
search, you will be prompted to exit or to continue the search with another
entry. In searches based on lists of several items, any items not found
will be noted.

The search technique used here will only locate the first item to match the
entry. This facilitates the speed of the operation and is the reason why
the entries should be unique.

Multi-Item Search

Option 5 of the Search Menu identifies records of a database file which
match criteria involving several features:

A. One or more items can be specified for search.
These are entered in appropriate field areas.
B. The multiplicity of acceptance can be specified.
The search logic can be set for ALL or ANY matches.

A full screen interactive entry display completely defines both the search
and presentation of the results. This screen is it is split into two parts.

The top half presents a full set of fields in which you simply enter the
items sought. The line for author search does not distinguish between
first and other authors; both fields will be searched (only the last name
is used for the first author). There are six separate entries available
under key words. These will be checked for an exact match in the key word
field of the database. However, as a special case, an exclamation point
preceeding a key word search entry will match a portion or subset of the
key word in the record. **This is very useful.**

The second part of the screen presents a series of questions to setup the
details of the operation. Enter 'Y' or 'N' or accept the default values
(shown in upper-case). Details defined are the type of search logic used
to match items, optional information that may be included in the output,
and the type of output (file, print, screen).

The first line of the setup section specifies how the word/phrase items will
be used in searching (abstract search is not available as of this writing).
The second line of the setup section defines the type of search logic and
will be discussed below. The third line determines the inclusion of key
words, abstracts, and sponsor in the search output. The forth line of the
setup directs output of the search to the screen only, the printer, or a new
database file for which a name will be requested.

In the event that a printed output is selected, there will be yet another
setup dialogue. This will allow specification of a header to be printed
with the output, the numbering of the output, how the first author's name
should appear, and the range of records over which to conduct the search
(if you should want to limit it).

The matter of search logic is important only if more than one search item
is entered. Search and match ALL items is equivalent to the Boolean AND
requirement that to be accepted a record must contain all of the criteria
items listed. Search and match ANY item is equivalent to the Boolean OR
requirement that to be accepted a record need only contain any one (or
more) of the criteria items listed.

The search logic in the primary setup screen is limited to one type and
cannot be mixed. It is possible to accomplish more complex, mixed mode
logic searches by selection of the OTHER option on the Search and Match
line. If you plan to use OTHER, skip over the top half of screen 4.5 --
fill out the bottom. You will then be presented with a new screen to set
up the complex search. Setup of the complex search requires building-up
the search string which contains three elements, the item sought, the item
field, and the logic. Examples are shown and there are several cross checks.

DOCUMENT GENERATION

Reports and topical lists based on reference material are addressed under
Document Generation, selection #5 from the Main Menu.

Reports can be prepared in either of two basic formats: a line (linear)
type used for most bibliographical listings, or a card (block) type for
preparation of `file cards' of several popular sizes. Either of these
reports can be directed to a hard copy printer or to a disk file that can
be used as input in other operations, e.g., a word processor.

Topical lists present special information based on single fields within the
records, e.g., subjects (key words), authors, sponsors, etc.

Reports

The two report options, #4 and #5, are essentially the same except for the
output: printer or disk file. Selection of either will prompt for

acceptance or modification of parameters in the report settings dialogue box.
This permits selection of the linear format or card format styles, and
selection of various options that may be included in the report. (Selecting
inclusion of abstracts will significantly slow the operation.)

On completion of the report settings, a file name will be requested if the
output is directed to a disk file. The disk option can be used in many
ways, including introduction of any needed embellishments, and inclusion of
lists in other documents. The disk file created is a DOS-Text (ASCII) type
which can be read directly into most word processors. It can be made with
or without a wordwrap (hard carriage return). Other types of disk file
generation are included under Export Utilities, Screen 2.4.4

If the output is directed to the printer, the report settings box will be
replaced by a printer settings dialogue for which the parameters can be
accepted or modified. Pitch, margins, indents, and page length can be set.
Note that the direct printing mode provides little in the way of character
embellishment e.g., underlining, italic, bold, subscripts. The printer
settings may have already been set in the Utility Menu (2.5), in which they
are explained in more detail. If not, they can be modified here. Most of
the parameters are self explanatory except for the printer codes which have
the following significance:
U = Universal printer E = Epson type printers
D = Diablo 630 printer T = Toshiba 351 printer
N = NEC 3550 printer L = HP Laser Jet printer
O = Other, defined by user in 2.5 Utility Menu

After completion of the various settings, a final screen of setup parameters
is presented which varies according to the format type being used. For the
linear format it provides for a report header, the type of numbering (if any)
to be used for the reported items, the author name sequence, and the range of
records to consider (except if indexed). For the card format the final setup
illustrates the card format and covers the size of card to be generated, the
author name sequence, the numbering, and the record range.

For the card format it is especially important to have chosen an appropriate
printer pitch. For example, using a 10 pitch font for a 3 x 5 card will work,
but will be crowded and use extra cards; 15 pitch (compressed) yields a much
better card form. Experimentation can be very useful for card printing.



Topical Lists

Topical lists present special information based on frequency of occurrence of
items within the database, e.g., subjects (key words), authors, sponsors, etc.
Selecting this option opens a dialogue box which provides for four types of
compilations and indicates their status. Operations options follow the type
selection and direct the output to the screen or printer. The operations also
include information regarding up-dating and compiling the particular list type.

Topical lists are reported in a two column format arranged alphabetically
(excluding leading left parentheses). These lists can be especially useful
in tracking the occurrence and consistency of key word usage. The authors,
source, and publication year can also provide interesting information.

The different types of lists are stored as special database and index files
in the data file directory. There is also a file which tracks the status
of these lists. The files used for these purposes have the same name as
the database file from which they are derived; the extensions differ
according to the type (*.D01,*.N01,*.MEM).


Supplemental Note: Abstracts print correctly to a printer, but cannot be
transferred to disk file in the formated Report output. This is a quirk of
the abstract (memo field) and will addressed in a later version when all
aspects of abstract handling are improved. Abstracts do transfer properly
as database file components, e.g., in searches.


UTILITIES

The utilities provide basic handling of the files and printer. They are
essentially self explanatory e.g., list, create, index, sort, import,
append, erase, etc. Only a few notes will be added here and below. A file
can be opened in any menu so it need not be done here. The printer set-up
can also be handled in the report menu (#5). Import/export functions, in
addition to handling foreign files, can be used to copy files for archival
storage. The delimited ASCII file takes up less space than the dBase form.

The file survey option (#2) provides listings of files that exist in a
particular data directory. These listings can be useful in the operations
of the organization option (#3) which include opening, indexing, sorting,
and erasing files. For this reason, the lists from option #2 can be
retained on the screen while moving to option #3. However, lists displayed
from the survey will not be updated by changes made in option #3; the
survey must be re-run to show the new status.


Note that there is a very important difference between the index and sort
operations. Both cause the database to be displayed in a particular order
based on the key field(s). Indexing does not alter the database file; it
generates a relatively small file that when activated, arranges the order
of display of database records. An active index will automatically be up-
dated as a database is changed.

Sorting generates a completely new database ordered as specified by the key
field. It requires disk space equal to that of the original database and
takes considerable time to generate. It is not automatically updated when
additions are made to the sorted database. Sorting is rarely of any great
utility.

Appending files provides a useful way to work with small files for
development purposes and then combine them into a master database file.
The append operation simply adds the records of a specified file to the
open, active database file. It also cleans up any extraneous material in
the memo (.DBT) file.

Import/Export utilities in option #4 provide very powerful tools for
manipulating data. They are intended to supplement import/export utilities
that are usually available in other software packages, a great many of
which support dBase type files to some degree. The utilities included here
are built from basic dBase type operations and syntax. They may require
some experimentation to achieve the desired result and it is recommended
that this be done with small test files before application to large files.

Import/Export selections will open up additional input requests and
specifications. There is a custom format that may be prepared by the user
and retained for subsequent use. This type will allow for definition of both
line and record delimiters and construction of complex line formats. The
custom format is built by example. [An extensive import/export translation
utility program is available to registered users.]

MISCELLANEOUS NOTES

There are various points to the operation of Reference Management System
that have not been treated under specific topics. They apply in several
areas or may have been beyond the scope of the general discussion of the
particular section. The following list is included to cover such items and
general comments.

1. In places where an index can be activated, the first input addresses
the choice to use the index:

Data base In Use:_______ Index? (y,N):___ Index Name:______

Accepting the default (y,N) without actually entering N will avoid
entry of an index name and go directly to the menu selection entry.
However, each time the menu reappears, the index question will have to
be answered again. On the other hand, if the response is to enter N,
then the question will be avoided the next time around. This can
speed things up if you do not intend to use an index. Also, typing a `?'
for the index name will bring up a list existing indes file names.

2. The Toshiba and NEC 3550 printer definitions have not been properly
incorporated and will not function. In the absence of any other
printer for which definitions are included, use default printer
settings or disk file transfer.

3. Abstract presentation requires extra time. This is because formatting
as stored in the memo field may not be appropriate. This process uses
a temporary disk file operation which is necessarily slow. The means
are now available so that this can be improved in a future version.

4. In the event of an error, a control window should be displayed to
allow some rational options other than crashing the system. The error
number and message will be shown and you should note these for
reference. There is a possibility to recover from the error by
pressing the space bar and this is always worth a try. The alternative
is to use the END or ESC keys to exit the program.

5. In regard to generation of index files (either permanent or temporary),
the field size is important. Do not attempt to form an index on the very
long fields. Fields shorter that 40 characters present no problem.
(See Screen 2.2.3 or the Appendix for lists of field lengths.)

6. There are a few places within Reference Managemant System where dBase
type commands or command fragments can be entered by the user. These
are generally conditions controlling the type and amount of material
returned on a search, report, or display. Examples have been given in
this manual, but may warrant some expansion. A typical example is of
the form:
'Jones' $ lst_name

where JONES is a string delimited with single quotes, $ is an operator
which interprets as 'is string contained in', and LST_NAME is a field
name to which the qualification JONES is applied. Other operators are
more obvious: =,<,>, and # (not equal to). Double quotes are also used
as delimiters.

Sometimes strings can be used in functions such as:

ASC(lst_name) > 71

where ASC() is a function that returns the numerical value of the ASCII
decimal code for the first character in the string. The string will be
taken from the field LST_NAME. The value 71 is the decimal code for the
upper case letter G. Therefore the qualifier illustrated will allow last
names starting with H or higher. This can be used for certain output
control, like starting over after an interruption.

Other functions of possible use are date and math functions; please refer
to the many good books on the dBase language for the proper use of these
and other functions.



APPENDIX



STRUCTURE OF THE MASTER DATABASE


Field Name Type Length

1 IDENT_NO C 5
2 INDEX_NO C 5
3 CODE_1 C 14
4 CODE_2 C 14
5 LST_NAME C 15
6 FST_NAME C 10
7 AUTHOR_2 C 64
8 SOURCE C 128
9 VOLUME C 10
10 YEAR C 10
11 PAGE_FRST C 6
12 PAGE_LAST C 6
13 TITLE C 128
14 KEY_WORDS C 128
15 COMMENT C 192
16 LANGUAGE C 10
17 TYPE C 11
18 NREFS C 5
19 SPONSOR C 32
20 ABSYN C 1
21 NOTES M 10


SPECIAL THANKS

I would like to mention Alan Simpson, who has no idea who I am. Mr.Simpson
has authored a number of excellent books on dBASE and dBASE program language.
His book "Understanding dBASE III", published by Sybex, was of great value to
me as a novice starting with dBASE III in 1985. I continue to refer back to
it, and its follow ups. I recommend them to anyone either starting or needing
reference material and suggestions on commands and programing.

I must also acknoledge the "Advanced Programmer's Guide, featuring dBaseII and
dBaseIII" by Luis Castro, Jay Hanson, and Tom Rettig, published by Ashton-Tate.
The opposite of Simpson's early books, this is a detailed reference on the
workings of dBASE. It is an excellent source of discussion on the intricacies
of the language.

Another reference that I found especially helpful in this project was "Command
Performance dBASE III" by Douglas Hergert, published by Microsoft Press. This
is an encyclopedic presentation of the command structure, including examples.

I must also mention Morrie Wilson of Wilson WindowWare, whose program for
Windows, COMMAND POST, convinced me that shareware can provide excellent
products, and can be advantageous to both author and consumer.

Last but not least, a personal thanks to my wife and family for patience during
the hours, days, weeks, etc. that I spent tied to the 'box' working on this
project through all of its ups and downs.


YET TO COME

As with any development, there is always room for improvements. These can
be based on discovery of bugs in the program, new improvements in the
development software, and user suggestions. Here are some of the changes
currently planned for future versions of Reference Management System:

1. Improved abstract handling. This is high on the priority list. It
will be possible to speed up handling of abstracts and to facilitate
searches within the abstract memo fields.

2. Check for duplication of records between two files will be added.

3. Conditional control of record report output will be added to allow user
specified condition at time of report generation.

4. Additional report output styles/formats will be added that can be
structured by the user to meet special requirements.

5. Special transfer formats (.FMT files) will be useful for moving record
data to and from DIALOGþ formats. (Now available as a separate utility.)

6. Control of firstname, lastname placement for all authors will be added.

7. More online help is intended, especially to access information between
the primary menus, e.g., index file names, key words, etc.

8. Improved coding to increase speed, provided better error trapping, and
reduce program size and memory use.