home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.amiga.misc
- Path: sparky!uunet!usc!cs.utexas.edu!uwm.edu!ux1.cso.uiuc.edu!usenet.ucs.indiana.edu!navajo!shulick
- From: shulick@navajo.ucs.indiana.edu (Sam Hulick)
- Subject: Good news! Tired of vague filenames? ...
- Message-ID: <BzC17K.n1v@usenet.ucs.indiana.edu>
- Originator: shulick@navajo
- Sender: news@usenet.ucs.indiana.edu (USENET News System)
- Nntp-Posting-Host: navajo.ucs.indiana.edu
- Reply-To: shulick@navajo.ucs.indiana.edu
- Organization: Vallen Software
- Date: Wed, 16 Dec 1992 03:18:55 GMT
- Lines: 55
-
-
- Hello, all.
-
- Maybe a few of you remember my old program, "Class." I think the latest
- release was v1.2. Anyway, it determined what type of a file something
- was. Well, it was crappy.. :) The source code was very sloppy and
- inefficient. So I'm doing a total rewrite, and this release will be
- TClass 3.0. It recognizes the following types of files:
-
- FORM files (IFF, that is. 'FORM' includes ANY type of IFF: 8SVX,
- ILBM, ANIM, etc.) Note, the way of determining this sort of
- clashes with the program's structure.. I'll worry about that later,
- though.
-
- Other things recognized: C source, Assembly source, *.o files, C
- includes, info files, executables, zoo archives, lzh archives, DMS
- archives, GIFs, fonts, MED songs, compressed files (*.Z), icons, MED
- modules, lha archives, MovieSetter anims, JPEGs, and uuencoded files.
- (whew) :) "Well, what about <filetype>!" you ask? :) TClass has a
- learn option, as well. (ALL of these filetypes are stored in
- S:TClass.brain) Anyway, let's say you (for some strange reason) wanted
- TClass to recognize MS-DOS executables. In the learning process, just
- as with humans, the more samples you have, the more accurate. For
- instance, if you just give LAMEPROG.EXE to TClass, it wouldn't have any
- basis of comparison. Now if you feed LAMEPROG.EXE, FOOBAR.EXE,
- BLECH.EXE, BIFF.EXE all into TClass, what it does is it compares Magic
- Numbers (tm) (fancy phrase for initial bytes) of all the files, and
- where it finds them common, it records these bytes as well as their
- positions, in S:TClass.brain. So, obviously, the more files of the same
- type you feed it, the more accurate future determinations will be. The
- structure of S:TClass.brain will be VERY strange. It will almost
- resemble source code. Here's a very rough draft, which will probably
- change once I get the project going:
- TClass.brain:
-
- MED module::0=77,1=77,2=68,3=48::
- Icon::0=-29,1=16,2=0,3=1,4=0,5=0,6=0,7=0,8=0::
- C source::@isext(.c)::
-
- Very simple. :) (the documentation will explain everything clearly)
- But for the most part, you won't have to edit S:TClass.brain, though you
- can, and quite easily w/ the help of another program I'll include
- (showbytes) and, of course, the docs.
-
- This will take quite some time for me to develop. I'm going on vacation
- soon, and when I get back, after I get situated with classes, I should
- have quite a bit of time to program this baby. Watch for it! :)
-
- Merry X-mas
-
- --
- // Amiga 3000 ___ \ Sam Hulick: shulick@indiana.edu (NeXTmail OK!)
- // 68030 25 MHz /__/\ \ My opinions wear combat boots. Or whatever.
- \\// OS 2.04 \__\/ / "Walk! Not bloody likely. I am going in a
- \/ / taxi." --George Bernard Shaw
-