Info-IBMPC@WSMR-SIMTEL20.ARMY.MIL ("Info-IBMPC Digest") (06/23/91)
Info-IBMPC Digest Sun, 23 Jun 91 Volume 91 : Issue 162
Today's Editor:
Gregory Hicks - Rota Spain <GHICKS@WSMR-Simtel20.Army.Mil>
Today's Topics:
Aproria Object Oriented Desktop help
BGI-Driver for Tseng ET-4000 wanted !
BoundsChecker SUMMARY!
Does IBM DOS 5.0 work on Clones?
Re: DOS 5 and memory management
IBM PS/2 311MB disk drive problem (V91 #128)
Index of the SIMTEL20 archives kept in mirrors/msdos
Re: MS-DOS 5.0 (V91 #145)
Large directories take time
Maxtor 213 and AWARD BIOS Incompatibility (V91 #143)
Today's Queries:
)V91 #132:0
HPGL
looking for '86 opcode-handling source code
looking for 2 programs
Mouse Driver patch needed
MS-DOS 5.0 Performance?
need database to support graphics
Send Replies or notes for publication to:
<INFO-IBMPC@WSMR-SIMTEL20.ARMY.MIL>
Send requests of an administrative nature (addition to, deletion from
the distribution list, et al) to:
<INFO-IBMPC-REQUEST@WSMR-SIMTEL20.ARMY.MIL>
Archives of past issues of the Info-IBMPC Digest are available by FTP
only from WSMR-SIMTEL20.ARMY.MIL in directory PD2:<ARCHIVES.IBMPC>.
----------------------------------------------------------------------
Date: 19 Jun 91 19:27:37 GMT
From: jaapv@accucx.cc.ruu.nl (Jaap Verhage)
Subject: Aproria Object Oriented Desktop help
In digest <V91 #146> Info-IBMPC@WSMR-SIMTEL20.ARMY.mil writes:
>Do you know of Aproria Object-oriented Desktop ?
>I learned it requires windows and it is on Compuserve (go PCMAG).
>I got this info from a friend and I will be grateful if you can enlighten
>me on that. Cause I have no access to Compuserve.
([gph's suggestions deleted])
[Thanks!]
The file is on cica.cica.indiana.edu, the current `Windows heaven'. As
far as I know, it's called `Aporia', not `Aproria'. Have a look in one
of the subdirs of /pub/pc/win3.
Regards, Jaap.
Jaap Verhage, Academic Computer Centre, State University at Utrecht, Holland.
jaapv@accucx.cc.ruu.nl +<-*|*->+ I claim *every*thing and speak for myself
------------------------------
Date: Tue, 18 Jun 91 21:05:13 PDT
From: 6500boo%ucsbuxa@hub.ucsb.edu (William Bushing)
Subject: BGI-Driver for Tseng ET-4000 wanted !
Carsten,
I use an ET-4000 based Diamond SpeedSTAR and found that the BGI driver
in SVGABG21.ZIP on GARBO.UWASA.FI (128.214.12.37) worked well with it
at higher resolutions in 256 colors but it did not support the mouse at
those resolutions so I was unable to use it.
Bill
William W. (Boo) Bushing
6500boo@ucsbuxa.ucsb.edu
Marine Biotechnology Lab
Univ. of Calif. Santa Barbara
------------------------------
Date: 18 Jun 91 22:35:47 GMT
From: sorrow@oak.circa.ufl.edu
Subject: BoundsChecker SUMMARY!
Comment: Forwarded by Keith Petersen
Well, here is the general review of BoundsChecker for DOS. For those
of you who don't know, BoundsChecker is a program that allows you find
nasty pointer problems. It requires the following:
386
256K extended memory
To be useful, IMHO, it should also be used with either TC or MSC (which
use the correct symbolic debug info...ie info that BoundsChecker
knows).
Of the replies I got, 4 were positive and 1 was not. I ordered it last
week and it arrived today. I haev had some time to play with it and
have found it to be an outstanding product for WHAT IT DOES.
The one dissenter amongst the responders stated that he believed Turbo
Debugger 386 was superior. As far as I am concerned, BoundsChcker is
not a generic debugger. You can't watch variables, etc. All you can
do is run your program and hope that somewhere you don't overwrite an
array boundary or part of your code. It is highly recommened for
software QA and finding nasty pointer bugs.
While using it today I must conclude that if you are a professional
programmer who uses C (and thus must deal with its wanton ability to
kill your programs) then this is a MUST. It will trap the following
bugs:
char a[80];
int i=0;
for (i=0;i<=80;i++)
a[i]=0;
It will also catch other things such as:
(char *)p=malloc(100);
*(p+101)=0;
Or executing a NULL isnrtuction, etc.
And the beauty of it all is that, yes, in a dos environment you will
have a problem with it at first due to all the LEGAL memory overwrites
you do. However, these can be ignored and placed in an exception file
so that it skips over them.
Bad pointers can be found in less than two minutes. It so far has
proven to be wonderful. As a debugger, it is hideous. As a bug
killer, it's great!
Nu-Mega Software makes it, along with a program called Soft-Ice and
CV/1. They seem to be knowledgable. You can find their add in Byte,
Dr. Dobbs, etc.
/*
Brian Hook -- MS-DOS Programmer for Contract
*/
------------------------------
Date: 17 Jun 91 19:20:05 GMT
From: woan@exeter.austin.ibm.com (Ronald S Woan)
Subject: Does IBM DOS 5.0 work on Clones?
Comment: Forwarded by Keith Petersen
Timmons@atccad.enet.dec.com writes:
>Thru July 10, Egghead has IBM PC-DOS 5.0 for only $39.99. That is
>cheaper than dealers pay for MS-DOS 5.0. But, the ad says for IBM
>PC's only! Is that true?
It is an upgrade that installs over a preinstalled IBM PC-DOS only.
The fully licensed version is $99 at Egghead, I believe. The generic
MSDOS 5.0 upgrade is also $39.99 anyway.
>To get the extra $20 off they ask you questions - could one of the
>questions be "What is your IBM-PC Serial Number?)
No, what speadsheet and wordprocessor do you use are the types of
questions they ask, along with do you want to buy the additional book
from Queue.
>Is this just a statment to be sure IBM does not have to respond to
>problems with clone? I can speculate, but has anybody actually done
>the install?
It may only be warranted for IBM machines... The full version should
work fine, except for BASICA which expects part of it in ROM.
All Views Expressed Are My Own And Are Not Necessarily Shared By My Employer
+ Ronald S. Woan
woan@cactus.org or
woan@austin.vnet.ibm.com +
+ other email addresses Prodigy: XTCR74A Compuserve: 73530,2537 +
------------------------------
Date: 18 Jun 91 16:30:02 GMT
From: jcmorris@mwunix.mitre.org (Joe Morris)
Subject: Re: DOS 5 and memory management
Comment: Forwarded by Keith Petersen
cb@tamarack12.timbuk (Chris Brewster) writes:
>Since the new DOS has paging and task switching, I'm unclear on
>what is still needed from a memory manager. Is QEMM or MAX still
>useful?
Both Microsoft and the memory manager vendors agree that what you get
with DOS 5 is about what the third-party vendors were offering four
years ago or so. The code packaged with DOS 5 does a pretty good job
of military packing; the other vendors (QEMM, 386^max, etc) do a better
job of grocery packing.
Military packing: load up the Herky bird until the next box won't fit,
close the cargo door, and dispatch. Anything which didn't make it on
the first pass is left behind.
Grocery packing (especially in a Volkswagen Bug): Put in all the
groceries which fit; if any are left over, empty the car and try again
until you find a combination in which every bag is inside the car and
there's still room for the driver.
Users who are really hurting for memory will still find the third-party
memory managers valuable, but much of their basic function is now part
of DOS.
Joe
------------------------------
Date: Wed, 19 Jun 91 09:54:23 GMT
From: a0045@rrz.Uni-Koeln.DE (Jochen Roderburg)
Subject: IBM PS/2 311MB disk drive problem (V91 #128)
Hello Bill,
As far as I understand it, the IBM *factory formatting* records some
information on a special sector on their disks which they need to
correctly handle the disk afterwards. They seem to need it already for
their *low-level formatting* (CTRL-A on the reference diskette).
Low-level formatting on another system destroys this information, and
you get the problem which you reported.
The only solution I know is to use a special PS/2 version of the
well-known OnTrack DiskManager Software which knows how to handle IBM
disks and is able to low-level format them the way IBM likes it. We
got this recommendation and the software from our computer vendor when
a while ago we were looking for a cheaper disk for a PS/2. Of course
he was happy to sell us a bit more than the naked disk, but he was
right, it was really needed. Without using it I made the same
experience with IBM's *low-level formatting* that you made. With
DiskManager I got it to work immediately.
The disk involved was a 345MB disk from Seagate/Wren/Imprimis, which
works without problems in a 3COM file server since then. I had the same
success a while later with a NEC 150Mb disk, which I now have as second
disk in my *personal workstation*, a PS/2 model 80-111 with AIX.
Jochen Roderburg
Regional Computing Center
University of Cologne
Robert-Koch-Str. 10 Tel.: +49-221/470-4564
D-5000 Koeln 41 Internet: Ro @ RRZ.Uni-Koeln.DE
Germany BITNET: A0045 @ DK0RRZK1
------------------------------
Date: Wed, 19 Jun 1991 11:56 MDT
From: Keith Petersen <w8sdz@WSMR-SIMTEL20.ARMY.MIL>
Subject: Index of the SIMTEL20 archives kept in mirrors/msdos
abdul%adler@ztivax.siemens.com (Abdulmesih Aktas) writes:
> I am looking for an Index of the SIMTEL20 archives kept in wuarchive's
> mirrors/msdos (something like the concatenation of all 00-index.txt
> files) ?
The file you are looking for is /mirrors/msdos/filedocs/simlist.arc
on wuarchive.wustl.edu and on SIMTEL20:
Directory PD1:<MSDOS.FILEDOCS>
Filename Type Length Date Description
==============================================
SIMLIST.ARC B 254607 910608 Text format list of all MSDOS files w/descrip.
It is a listing of the entire MS-DOS collection in the same format as
the 00-INDEX.TXT files.
Keith
--
Keith Petersen
Maintainer of the MSDOS, MISC and CP/M archives at SIMTEL20 [192.88.110.20]
Internet: w8sdz@WSMR-SIMTEL20.Army.Mil or w8sdz@vela.acs.oakland.edu
Uucp: uunet!wsmr-simtel20.army.mil!w8sdz BITNET: w8sdz@OAKLAND
------------------------------
Date: Wed, 19 Jun 1991 09:45:19 PDT
From: Nina_J._Thayer.ES_CP8@xerox.com
Subject: Re: MS-DOS 5.0 (V91 #145)
Loaded it over the weekend, and got an instant rebate of $20 at Egghead
(no mail in stuff).
Have a 386SX with 640 and extended memory to 1 MB, 124 MB hard drive, 3
1/2 and 5 1/4 high density floppy drives. System 5.0 seems to work
fine with Windows 3.0, and assorted Window and other applications, uses
50K less RAM on my system than did 4.01 or 3.3, and lets me open
various applications within Windows that I had trouble running
before.due to memory messages.
At least this is more compatible than my System 7.0 installed on my
Macintosh, which DOES have some compatibility problems. It seems that
DOS is DOS, whether 3.3 or 5.0, but System 7.0 for the Mac is a new
breed. NT
------------------------------
Date: Wed, 19 Jun 91 11:29:22 EDT
From: pilant@NADC.NADC.NAVY.MIL (J. Pilant)
Subject: Large directories take time
"Gisbert W.Selke" <S00100%DBNRHRZ1.BITNET@CUNYVM.CUNY.EDU> writes:
>We are currently building an application that seemingly requires us to
>have a directory filled that, in its turn, contains *many*
>subdirectories (1500+), each of which contains 3 files which are to be
>processed by SPSS. The access times drastically increase with the
>relative position of the subdirectory within the directory...
> [table of dir-position vs time was here]
>Now, there are some regularities here. Times increase not continuously
>but in discernible steps. Within each of the groups thus defined, there
>are several clearly discernible strands. And there are runaway times
>near to the beginning of each of the groups.
>
>Now, what are the reasons for these effects?
Each DOS directory entry takes 32 bytes. This gives 16 entries
per sector. On a 32MB disk, a cluster is 2KB = 4 sectors = 64 entries.
DOS reads the directory from the top and must read each sector as it
goes. The DOS BUFFERS will do some cache effects, but it may not be
sufficient.
On a full (1500+ entry) directory, you need to read an average of
over 45 sectors just to find where the file/subdir starts on disk. at
4 sectors per cluster that is 12 clusters average access multiplied by,
say a 28ms disk, the disk access time and you get an average acces of
336ms to find the file/subdir start. Compare this to a small, less
than 64 entry, directry. There will be 1 cluster and therefore 28ms
access if it is not in the BUFFER already. To open each entry in the
1500+ directory will require a grant total of 504 seconds just of
access time.
>And, more importantly, is there a (standard DOS) way around this
>tremendous increase in time?
A simple way around this is to avoid a long directory. As you
stated that the structure was not important, why not define your
directory as:
old: ...\1234\...
new: ...\12\34\...
or: ...\12\3\4\...
this two (three) level split will have fewer entries in each dir. for
the 1500+ sequence, there will be 15 (28ms) + 100 (28/56ms) = 56/84ms
access with 2-8 sectors that may be in the BUFFERS already. With the
three level sequence, it will be 15 (28ms) + 10 (28ms) + 10 (28ms) =
84ms again with BUFFERS covering the 3-4 sectors needed.
/jeff
------------------------------
Date: 19 Jun 91 16:20:00 EDT
From: "Raymond P. Pichulo" <pichulor@v3.hanscom.af.mil>
Subject: Maxtor 213 and AWARD BIOS Incompatibility
>A friend bought a Maxtor 213 ( 200 megs ) drive from a flea market.
>This is an IDE drive. When he plugged it in his AWARD BIOS machine he
>only managed to get 130 Megs out of 200. We know that our BIOS does not
>support 26 sector drives, that means we need a low level formatter with
>a device driver to do the job. We brought the machine to a local store
>so that they would use DM (disk manager) on it, but DM did not have
>the 213 drive , so the folks gave up. A third party low-level formatter
>we used only allowed 17 sector formats (MFM probably), so there goes
>that idea.
A couple of things... First, it is possible that the local store
probably doesn't know how to use Disk Manager. Besides the automatic
mode, there is a manual mode in which you can enter the drive
parameters for one that DM doesn't know about. However, this takes
time and the whole process takes longer. Even easier, though, within
DM there is an ASCII file called DRVPARMS.OCS (or something close to
it) that contains all the drive info. The file could be edited to add
the parameters for your Maxtor. Probably academic at this point since
you don't have DM and the store may not be willing to muck with it.
Anyway, that's one possibility.
Second possibility: You didn't mention what type machine (286, 386)..
I am not up on all the machine and BIOS combos, but my 386 has AMI
BIOS's in it and it has provisions for user-specified drive
parameters.. very convenient. It also has a routine to low-level
format the hard drive(s), so it sounds as if that may work. Possibly
you can get an AMI or other BIOS that has that capability. One
possibility is a company called Upgrades, Etc. in Redmond WA (800
541-1943) who sell BIOS's and seem to be knowlegeable in such matters.
I bought an upgrade BIOS ROM for an AT from them for $69.95.. It also
allows user defined drive spec's, low level formatting, etc. That may
blow your friend's budget, but is a second alternative.
Third alternative: You can buy your own copy of Disk Manager. I
forgot the name of the company that makes it, but it costs about $110
retail. This is a generic version which allows you to specify any
disk. They provide light versions of DM to the disk drive manufacturer
which are specific only for them. I never had a drive-specific copy to
muck with (maybe DRVPARMS file is all that needs to be changed to make
it work).
I don't know if this helps... everything I suggested costs money :^)
but good luck.
Ray
------------------------------
Subject: Today's Queries:
Date: Wed, 19 Jun 91 16:42:16 EST
From: Uhrmacher Mark <Uhrmacher_Mark@Bethesda.BAH.pc.niaid.nih.gov>
Subject: HPGL
Does anyone out there know where I can get a complete listing of the
commands in HPGL language??
- Thanks -
___
(|)ark
------------------------------
Date: Wed, 19 Jun 91 15:55:51 EDT
From: eli@smectos.gang.umass.edu (Eli Brandt)
Subject: looking for '86 opcode-handling source code
I'm writing some code which needs a "mini-debugger" as part of the
application. I'd like to find source code to disassemble '86 machine
code into something my program can comprehend. Towards that end, does
anybody have source code for an assembler, a disassembler, a debugger,
an emulator, or a simulator?
Eli Brandt eli@gang.umass.edu
------------------------------
Date: 19 Jun 91 15:04:55 CDT (Wed)
From: b11!bill@uunet.UU.NET (Bill Welch)
Subject: looking for 2 programs
Summary: Reposted by Keith Petersen
I am looking for a couple of dos/pc programs:
1) something that will redirect AUX output to a monichrome card rather
than the serial port.
2) something that will intentionally write a "bad sector" (crc error)
to a floppy diskette. i am evaluating the robustness of a particular
copy protection scheme, and i am pretty sure that it can be defeated if
a way to write bad sectors is widely known/available. please note that
i am not trying to cheat anyone, but am rather just testing how good
this scheme is.
thanks in advance for any help!
-bill welch ab4yd
bill@ingr.com
------------------------------
Date: Wed, 19 Jun 91 10:38:15 EDT
From: pgauvin@ncs.dnd.ca (Pierre Gauvin)
Subject: Mouse Driver patch needed
I am working with somebody with cerebral palsy. While he can usually
handle a mouse reasonably well, he finds it difficult to double click
without moving the mouse. Since he likes Windows, I wondered if there
is a way to replace the double click with some other key combo.
What I thought I would do is try patch the mouse driver so that , say,
ALT-left button would be seen as a double click. I looked in whatever
info I have on mice and interrupts and could not find anything
relevant. Any suggestion?
I am looking at both Microsoft and Logitech mouse drivers.
Thanks
Pierre Gauvin
pgauvin@ncs.dnd.ca
------------------------------
Date: Wed, 19 Jun 91 12:45:10 EDT
From: mbeck@ai.mit.edu (Mark Becker)
Subject: MS-DOS 5.0 Performance?
There have been many messages about 5.0 and praises over regaining 40K
to 50K of the low memory space while running on 80x86 class machines.
Has anyone benchmarked the performance between 3.3, 4.01, and 5.0 yet?
My interests are along the lines of decreasing disk access time
relative to the DOS filesystem and applications programs. A couple of
days ago I downloaded a copy of THRASHER.ZIP (available from a local
BBS) and ran it against DOS 3.21 and 3.3. 3.21 was nearly nine seconds
faster at dealing with the disk than 3.3 for nearly all values of
BUFFERS in the CONFIG.SYS file.
Regards,
Mark Becker
Internet: mbeck@ai.mit.edu
------------------------------
Date: 19 Jun 91 9:11 -0800
From: Rick Gee <gee@admin.okanagan.bc.ca>
Subject: need database to support graphics
i will be getting involved with a project to collect data on natural
areas (ponds, forests, etc.) in the local area. can anyone suggest a
database that will allow
a. essentially unlimited-length text fields
b. graphic images, e.g maps or photographs to be stored.
------------------------------
End of Info-IBMPC Digest V91 #162
*********************************
-------