home *** CD-ROM | disk | FTP | other *** search
-
- PC-KAT (TM)
-
- The Easy, Practical, Super
- Capacity Diskette Cataloger
-
-
- TECHNICAL REFERENCE MANUAL
-
-
- by
- R.H. Martin
-
-
- Applied Foresight (TM)
- P.O. Box 20607
- Bloomington, Minn. 55420
-
- January 17, 1986
- Version 1.2
-
-
-
- Copyright 1986 by R.H. Martin
- All Rights Reserved
-
-
-
- A DB-SHARE Software Product
-
-
-
-
-
-
- A limited license is hereby granted to copy
- and distribute wholly unmodified copies of the
- PC-KAT Software Diskette which contains this manual
-
- (Read the manual, "THE DB-SHARE PROGRAM"
- for full details about this limited license)
-
-
-
-
-
-
- Trademark Pending
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- TECHNICAL REFERENCE MANUAL Copyright 1986 R.H. Martin Page 2
-
-
-
- TABLE OF CONTENTS
-
- Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
- About This Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
- Using PC-KAT With Tall Tree Systems' JET Program. . . . . . . . . . . . 2
- Efficiency Of PC-KAT . . . . . . . . . .. . . . . . . . . . . . . . . . 2
- PC-KAT Inquiry Search Strategies. . . . . . . . . . . . . . . . . . . . 4
- Recovering Your Database. . . . . . . . . . . . . . . . . . . . . . . . 5
-
-
- ABOUT THIS MANUAL
-
- With the exception of the Registration Benefits section, this manual
- restricts itself to the base PC-KAT software diskette. The concept of
- "base software" is discussed in greater detail in the THE DB-SHARE
- PROGRAM manual.
-
- A section is designated by a centered and underlined heading. Centered
- but NOT underlined headings are referred to as a subsection.
-
-
- USING PC-KAT WITH TALL
- TREE SYSTEMS' JET PROGRAM
-
- PC-KAT also catalogs, without requiring any data entry, the subdirectory
- which a diskette file resides in. This is a feature that many owners of
- the high speed copy utility, JET by Tall Tree Systems, will appreciate.
- JET is able to copy files to diskettes and preserved their subdirectory
- location on the targeted (diskette) volume. Many IBM AT owners may
- appreciate this PC-KAT feature too. With the 1.2 MB capacity floppies,
- there is increasing use of making subdirectories on floppies. You do not
- need JET nor an IBM AT to make subdirectories on your floppies. Just
- type: MD A:\SUBDIR and press ENTER to create the directory called,
- \SUBDIR, on drive A.
-
- JET is one of the programs available on the JETDRIVE diskette, which
- includes a super RAM Disk software for the IBM PC and compatibles.
- JETDRIVE retails for $60 and is available by Tall Tree Systems, 1032
- Elwell Court, Suite 124, Palo Alto, CA 94303. Some mail order houses,
- like Applied Computer Products, Inc., 1-800-223-3860, sell JETDRIVE for
- as little as $40.
-
-
- EFFICIENCY OF PC-KAT
-
- PC-KAT is fast and efficient, minimizing the amount of disk space
- required to store the cataloged information. Each date and time field is
- concatenated together and stored as a 4 byte long integer. In plain
- words, what you see on the screen as 10 bytes of information is stored in
- 4 bytes of storage. The four byte volume ID is stored as a 2 byte
- integer. If a diskette filename is associated with a floppy
-
-
-
-
-
-
-
-
-
-
-
- TECHNICAL REFERENCE MANUAL Copyright 1986 R.H. Martin Page 3
-
-
- subdirectory, the directory name is separately stored in the database
- only once in its alphabetical description form, and given an ID number.
- This ID number is carried in the appropriate individual filename records.
-
- The entire PC-KAT database is kept in the files prefixed as KATDB.DA and
- KATDB.KE. The database contains the system control record, the diskette
- volume text information and the directory file information for all
- cataloged files. In addition, all the record pointers are stored in the
- database to allow quick updating and retrieval of information. The file
- KATDB.DBD is the data base structure dictionary which never grows but is
- needed for the programs to work and must reside in the same directory
- area on the hard disk alongside the other KATDB files.
-
- What this means is that although PC-KAT must keep its inter-relationships
- connected at the disk drive level versus smaller packages which do most
- of their inter-relationships at runtime in memory, PC-KAT does not use up
- hardly any more disk space, and that is because of its efficient use of
- disk space. Additionally, PC-KAT has to allow for its database pointers
- to expand up to 16 million records. What this means is that on the
- average, PC-KAT uses up only 7 percent more space per cataloged directory
- entry than does its competitors.
-
- The other packages gain their processing efficiency by storing the entire
- database in memory at run time. That is why they can't get larger,
- because they are constrained by available program data memory space.
- Yet, without significantly sacrificing additional disk space, PC-KAT uses
- disk space to hold its pointers, and achieves comparable processing
- efficiency on a catalog database capable of holding 16 million records.
-
- The author maintains for himself a PC-KAT database of over 12,800
- filenames and around six hundred diskettes.
-
- Throughout the built-up of this database, the author has not noticed a
- worsening of processing performance, except that a display request for a
- floating substring phrase within any and all cataloged filenames does
- take a little longer as the database grows. To use the DOS DIR command
- syntax to explain this type of search request, although DOS does not
- honor such a search request, if it did, a request might look like this:
- DIR *CAT*. On a COMPAQ PLUS, with 8,000 filenames in a PC-KAT database,
- it takes under 2 minutes for such a search to finish, whereas with 6,000
- filenames, it takes under 1.5 minutes. Search requests for left-hand
- anchored substring phrases within any and all cataloged filenames is
- extremely rapid, taking only a couple of seconds to fulfill. A search
- request at the DOS DIR level might look like: DIR CAT*.
-
- PC-KAT is fast because of the tools used in developing PC-KAT (discussed
- earlier) and because of its streamlined design. For instance, if you
- have a diskette previously cataloged by PC-KAT, when you change, add or
- delete a few files on the diskette, PC-KAT concentrates on these
- modifications. A less efficient design might start out the processing of
- each diskette with the deletion of all prior database information for
- that diskette. Because of PC-KAT's usage of the sophisticated DB-VISTA
-
-
-
-
-
-
-
-
-
-
-
- TECHNICAL REFERENCE MANUAL Copyright 1986 R.H. Martin Page 4
-
-
- database manager, it is able to immediately access prior information for
- the at-hand diskette in the correct processing order in which it needs
- it. Finally, PC-KAT is fast because it allows the user to use a RAM disk
- emulator to store the subprograms and work files for PC-KAT on the RAM
- disk.
-
- PC-KAT is also fast because it is smart. If you mistakenly insert a
- diskette which hasn't changed since it was last processed by PC-KAT, you
- are told this is the case. PC-KAT keeps for each diskette both a numeric
- total of the number of files on the diskette plus the sum of all dates +
- all times + all file sizes of all the files on the diskette. Before PC-
- KAT considers updating directory entries for a diskette, it compares
- these two new totals with the prior run's old totals.
-
-
- PC-KAT INQUIRY SEARCH STRATEGIES
-
- To understand the many ways you can search your PC-KAT database and how
- efficient will be your search, you need to know how your PC-KAT database
- is structured. For searching purposes, you can think of the database
- being ordered in two different ways:
-
- (a) by alphabetical, indexed, order by FILENAME.EXT, regardless of the
- volume ID; and
- (b) by numeric, indexed, order by volume ID, and within volume ID order,
- subordered alphabetically, by FILENAME.EXT.
-
- The path information is carried in the database, but it has no ordering
- effect in terms of aiding retrieval of information.
-
- Based upon the structure of PC-KAT's database design, the following
- examples represent a range of searchs ordered by a most efficient search
- argument to a least efficient search argument:
-
- FILENAME ARGUMENT TYPE EXT ARGUMENT TYPE PATH ARGUMENT TYPE
- -------- - --- - -------- -
- MYFILE E DOC E MYPATH E
- MYFILE E DOC E MY A
- MYFILE E DOC E AT U
- MYFILE E D A <left blank>
- MYFILE E O U <left blank>
- MYFILE E <left blank> <left blank>
- MY A DOC E MYPATH E
- MY A DOC E MY A
- MY A DOC E AT U
- MY A D A <left blank>
- MY A O U <left blank>
- MY A <left blank> <left blank>
- FIL U DOC E MYPATH E
- FIL U DOC E MY A
- FIL U DOC E AT U
- FIL U D A <left blank>
-
-
-
-
-
-
-
-
-
-
-
- TECHNICAL REFERENCE MANUAL Copyright 1986 R.H. Martin Page 5
-
-
- FIL U O U <left blank>
- * FIL U <left blank> <left blank>
- <left blank> DOC E MYPATH E
- <left blank> DOC E MY A
- <left blank> DOC E AT U
- <left blank> D A <left blank>
- <left blank> O U <left blank>
- <left blank> <left blank> MYPATH E
- <left blank> <left blank> MY A
- <left blank> <left blank> AT U
-
- Don't be afraid to use the search strategy you need to accomplish the job
- you want done. For instance, against a database of 12,000 directory
- entries, to search using an argument like the above line with an asterick
- preceeding it, took around two and a half minutes. This search argument
- causes the entire database to be scanned from start to end.
-
-
- RECOVERING YOUR DATABASE
-
- "Into each life a little rain must fall." It can happen to one's
- database too. Recovery should be done when you loose your hard disk or
- when something has damaged your database. In either case, the database
- recovery procedure is the same. True, you have to restore everything
- else in a hard disk failure, but you're on your own to figure that
- dilemna out. Just the database recovery procedure is documented here.
- Follow these procedures.
-
- # ACTION EVENT OR COMMAND COMMENT
- -- ------ ---------------------------------- -----------------------
- 1 Type C: Move to your hard disk.
-
- 2 Type CD \ Move to the root
- directory.
-
- 3 Type VERIFY ON Ensure better file
- copies.
-
- 4 Type COPY KATDB.LOG KATDB.FIX Save current log file.
-
- 5 Type KATLOG Print your current log
- file.
-
- 6 Find Latest PC-KAT database backup disks Insert disk 1 into
- drive.
-
- 7 Type RESTORE A: C:KATDB.* Restore old PC-KAT
- database.
-
- 8 Do PC-KAT Integrity Analysis if not If integrity fails go
- using a "VERIFIED" backup set. back to step 5.
-
-
-
-
-
-
-
-
-
-
-
-
- TECHNICAL REFERENCE MANUAL Copyright 1986 R.H. Martin Page 6
-
-
- 9 Read KATLOG printout for line # of B/U state which you're restored
- to. Call this number "RESTORE-LINENUM".
-
- 10 Read KATLOG printout for last line # of printout. Call this
- number "LAST-LINENUM".
-
- 11 Calc Write down what LAST-LINENUM minus RESTORE-LINENUM equals.
- Call this number STRIP-LINENUM.
-
- 12 Type KATSTRIP STRIP-LINENUM Substitute real number.
- KATDB.FIX now contains
- an unsorted list of
- disks to reprocess.
-
- 13 Type SORT <KATDB.FIX >PRN This printout is sorted
- by yet-to-redo volume
- ID within location.
-
- 14 DO KAT -U or some variation Reprocess listed
- diskettes.
-
- 15 Type DEL KATDB.FIX No longer needed.
-
- You're done! Consider doing a PC-KAT Integrity Analysis and Backup to
- preserve the recovery work you just went through.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-