Info-IBMPC@C.ISI.EDU.UUCP (02/26/87)
Info-IBMPC Digest Wednesday, 25 February 1987 Volume 6 : Issue 12
This Week's Editor: Billy Brackenridge
Today's Topics:
FASTCARD EMS Board
3COM
re-ctty
3.5 Drives (2 Msgs)
Format 3.5 Drives
TURBO C (2 Msgs)
DisplayWrite3 and ASCII files
DisplayWrite3 and ASCII files
PC To Apple Disk transfer
MSC Floating Point Bug?
Bug in Microsoft MASM or DOS 3.2 or RAMdisk drivers?
SWPTR.COM Swaps LPT1 & LPT2
MSC 4.0 "system" function
LA-50 on the Serial Port
Another Bug in Microsoft MASM 4.0 (2 msgs)
Today's Queries:
Diagnostics
KISS Laser Software
SWPKEYAT.ASM
Copying 0 length files
Disk optimizers
Wierd problem with Mace Utilities disk
Buffers
Use 2 8250's to get 1200/75
Herc Plus SIMCGA Compatibility
Btrieve
MS Fortran Version 4.0
MS-DOS Using Extended Memory
ARC51.COM DOWNLOAD PROBLEMS
MINIX dl (2 msgs)
Reentrant ROMable 8087 Emulator Supplier?
PMAP VER. 1.1
Clock Calendars
From Addresses
Amstrad PC-1512
EMM for PC-elevATor
Program to Edit Binary Files (2 msgs)
----------------------------------------------------------------------
Date: Wed, 18 Feb 87 09:49:22 est
From: decvax!ittatc!ittvax!jeffd@ucbvax.Berkeley.EDU (Jeff Denenberg)
Subject: FASTCARD EMS Board
In reply to requests for information on "low cost" EMS boards:
I have used several Thesys FASTCARD IV EMS cards with good success. The PMI
Fastcard is the same device sold through a third party at ~$300. There are two
versions:
Fastcard IV - 2 MB EMS, serial port, parallel port, clock, and game port
Fastcard III - Only 2 MB EMS, up to 3 can be used with a Fastcard IV to
provide up to 8 MB of EMS memory.
The software and documentation provided is good. The only problem I know of is
that their EMS disk cacheing software (which is quite good) cannot handle some
newer disk drives (eg:RLL encoded with 25 sectors per track).
------------------------------
Date: Mon, 23 Feb 87 00:06:50 est
From: jpm@bnl.arpa (John McNamee)
Subject: 3COM
I've been unhappy with 3com software ever since I used their TCP/IP stuff
on a PDP-11. Their hardware works just fine, but they just don't have it
together in the software department.
I wouldn't think of doing a PC network with anything other than Novell
Netware. Netware has always offered the highest performance and richest set
of features of any PC networking package, and it runs on just about all the
network cards available for the PC (it will even gateway between different
physical networks). It is expensive, but so is wasting time with junk like
the 3com software.
John McNamee <jpm@BNL.ARPA>
(Usual disclaimer applies: I have no connection with Novell other than as
a happy customer)
------------------------------
Date: Mon, 23 Feb 87 06:25:52 EST
From: David Kirschbaum <kirsch@braggvax.arpa>
Subject: re-ctty
Re the query in INFO-IBMPC Vol 10 (couldn't make the return address
work) on CTTY:
Fought a long hard fight, and finally made it work. The tricks are:
Yes, use MODE to set the COM1 or COM2 port before you CTTY.
A full 25-pin cable does no good. The secret is a NULL MODEM cable.
Yes, as I recall, I did have to switch pins 2 and 3.
Worked just fine, running my XT clone (PC-DOS 3.1) from a Liberty
Electronics Freedom 100 terminal. (Well, fine for command line stuff.
Any program doing anything other than ANSI didn't do so well!)
Regards,
David Kirschbaum
Toad Hall
kirsch@braggvax.ARPA
------------------------------
Date: Wed, 25 Feb 87 20:17:26 EST
From: "termmaint" <acn@j.cc.purdue.edu>
Subject: 3.5 Drives
Regarding 3.5 inch drives...In order to be able to format 3.5 inch
drives at 720KB, you must use the latest version of Setup for the
IBM AT. It will support the 3.5 inch drive. IBM makes no mention
of this in the accompanying documentation, they simply believe
that everyone has the latest version of Setup :-)
So, to utilize 3.5 inch drives, use the new Setup.
Steve Gerber
acn@j.cc.purdue.EDU (arpa)
FidoNet Opus 201/1 317/423-2281 2400 baud
------------------------------
Date: Wed, 25 Feb 87 09:58:07 pst
From: well!nsmith@lll-lcc.ARPA (Neil Smith)
Subject: 3.5 Drives
In response to Lawrence <LDURHAM%SHEDIR@usc-oberon.ARPA>'s question
concerning 720K drives in DOS 3.2:
I have a standard 5.25" drive (A:) and a 3.5" 720 KByte drive (B:).
Running MS-DOS 3.2, I don't use the driver.sys device driver at all. The
documentation states that this is for EXTERNAL floppy drives (which may
be why it calls your new drive D:). Rather, I use the 'drivparm' command
in config.sys. This is documented on page 251 of the MS-DOS 3.2 User's
Reference (in appendix B). To get MS-DOS 3.2 to recognize a 3.5" floppy
as drive b B:, try:
drivparm=/d:1 /f:2
This tells MS-DOS that drive B: (/d:1) is a 720 KByte drive (/f:2). Do this
before adding any device drivers, as the documentation says that setting
'drivparm' will override any previous block device driver definitions. There
are many other options dealing with sectors/track, tracks/side, etc.
I have no problems accessing my 3.5" drive, formatting, etc. with this
approach. Perhaps PC-DOS 3.2 doesn't have (or has, but doesn't document)
this feature? It was fairly easy to find in the MS-DOS documentation.
-- neil smith ...!ihnp4!lll-lcc!well!nsmith
"DBF!!"
------------------------------
Date: Mon, 23 Feb 87 10:16:14 EST
From: Chris Schmandt <geek@MEDIA-LAB.MEDIA.MIT.EDU>
Subject: Format 3.5 Drives
To format the internal 3.5 inch drive, you have to access it as d:.
Once formatted, you can read/write off it as either b: or d:.
I don't know why, but the doc that came with mine told me the above.
chris
------------------------------
Date: Mon, 23 Feb 87 10:39:20 PLT
From: Don Howes <HOWESDW%WSUVM1.BITNET@wiscvm.wisc.edu>
Subject: Turbo C
To: INFO-IBMPC@USC-ISIB.ARPA
In response to the query about Borland's Turbo C, this product has not
yet been released. I spoke with Borland last Friday (Feb. 20) and they
gave me a release date of the second quarter of '87. Since the compiler
is not yet in beta test, they may or may not make this.
I requested a beta test copy of the compiler (I write a C programming
column for a magazine) for evaluation, but the person I spoke to could
not make me any assurances. If I do get one, I'll post my observations
to the net.
------------------------------
Date: Wed, 25 Feb 87 07:08:49 CST
From: mlw@ncsc.ARPA (Williams)
Subject: TURBO C
I picked up the following PRESS RELEASE (not review!!) about Turbo C.
There seems to be a lot of interest, so I thought I'd pass it on.
This comes from PROTECH, one of our public bbs brethren that's oriented
toward C.
FOR IMMEDIATE RELEASE
January 26, 1987
BORLAND CHALLENGES ALL PLAYERS WITH TURBO C: OFFERS FASTEST OPTIMIZING C
COMPILER FOR MICROCOMPUTERS AT $99.95
Programmers Can Compile From Command Line Interface Or From Borland's
Integrated Environment
SCOTTS VALLEY, Calif. -- Programmers embarking on serious
software development projects, and novices taking their first
crack at the C programming language, can harness the power of new
Turbo C from Borland International, Inc. The new language
development system compiles at a fast 7000 lines of code per
minute. Turbo C is the only optimizing C compiler for
microcomputers that provides end users with a choice of using
either a conventional-style command line interface; or the
integrated environment that Borland has become known for,
complete with built-in editor, compiler and Linker.
Turbo C provides a one-pass compiler, full control of memory
models, and extensive code optimizations. In benchmark tests
(attached), Turbo C generates the fastest, smallest object code.
Turbo C has a built-in Turbo linker, also provided as a
standalone program for use in the command line mode.
According to Borland president Philippe Kahn, "People use C
because of the conciseness and flexibility of the language...and
Turbo C is best described as a high level language tool that
delivers programming control at the low end as well. We are proud
to repeat the concept of Turbo Pascal, by delivering a lightning
fast compiler and a feature-rich development environment at a
fraction of the cost of what might be considered as a comparable
offering."
Turbo C's Compiler and Linker
Turbo C is a single-pass compiler that generates intermediate
data structures in memory. In contrast, most C compilers are four
or five-pass compilers that compile using temporary files on
disk. Turbo C compiles to memory at fast raw compile times
exceeding 7000 lines/minute on a 6 MHz IBM PC AT.
Turbo C supports 6 memory models: Tiny, Small, Compact, Medium,
Large, Huge. Near pointers and far pointers allow the programmer
to take full advantage of the 8086 microprocessor's architecture
with the mixed model technique employed in Turbo C.
A fast linker, like that implemented in Turbo Prolog, is
compatible with the DOS linker and links approximately two to ten
times faster than the DOS linker while it is about one-fifth the
size.
Turbo C features a built-in LINT with extensive error checking
for streamlined development and debugging; and full support of
ANSI prototypes. Optional Pascal calling conventions have been
included for added efficiency and optimization power.
A full range of Compiler options are provided including inline
assembler, multiple levels of optimization, generation of
80186/80286/8087 instructions, warning supression, and multiple
memory models. Compiler optimizations include automatic register
assignment and common sub-expression elimination.
In Support of Industry Standards for C
and Runtime Libraries
Turbo C implements the forthcoming ANSI C standard and delivers
full support of Kernighan and Ritchie C. Special extensions for
the PC environment include six memory models; and extensions for
mixed-language, mixed-model programming. Support for software
interrupt functions, and unique support for inline assembler, are
both implemented in Turbo C.
The Turbo C library includes standard functions and support for
the IEEE Floating Point standard (with software emulation if no
8087 is present). Support for UNIX compatibile library routines,
and supplemental functions for DOS and BIOS calls are also
included in the library for Turbo C.
The Development Environment
Turbo C features a pull down menu environment that is the
hallmark of other widely-used Borland language tools such as
Turbo Prolog and the forthcoming Turbo Basic. Because Turbo C
also comes complete with a command line version, professional C
programmers and software developers can select compile options
from the command line, from the integrated environment or from a
configuration file; without the need to set environment
variables.
The integrated environment in Turbo C supports an automatic
Make/Project Facility designed for managing program development.
A separate Make program is provided for the command line version.
This Facility compiles all modules in a project, or recompiles
only the modules that have been modified since last compiled.
The Run operation within the integrated environment will
recompile all necessary files, generate the executable code and
run the program, then return back to the user interface (UI).
Two tiled, zoomable windows -- the Editor Window and the Message
Window -- can be invoked within the editor; and the user can
toggle between these windows with a hot key.
In the Editor Window, a full screen editor is provided with
insert/overwrite, auto-indent, and block copy, move, read, write,
and delete along with many other functions.
In the Message Window, warnings and compiler error messages can
be turned off selectively at different levels. Turbo C steps
through multiple errors, and the interactive editor automatically
positions the cursor in the source code at the point of error.
The integrated help system in Turbo C is an abbreviated form of
the reference manual, showing declarations and descriptions of
all library functions for the large reference libraries commonly
associated with the C language.
The Turbo C User Manual
Turbo C's user manual is designed with special sections for both
experienced programmers, and for novice C programmers. The manual
contains a tutorial section on Turbo C for Turbo Pascal users
learning to program in C.
Borland International, Inc., was founded as a company devoted to
the development of superior language tools. The company's major
language development software includes Turbo Pascal, Turbo
Prolog, Turbo Basic and now Turbo C, and their accompanying
Toolbox products. Turbo Pascal and Turbo Prolog are the leading
language tools in their respective categories. Both Turbo Basic
and Turbo C are scheduled for delivery beginning in the first
half of 1987.
Turbo C will be available for $99.95 (suggested retail price)
either directly from Borland, or through its distributors and
dealers worldwide. For more information, contact Borland
International, 4585 Scotts Valley Drive, Scotts Valley, CA,
95066. Telephone (408) 438-8400.
Benchmarks
----------
Sieve benchmark (25 iterations)
Turbo C Microsoft C
Compile Time 3.89 16.37
Compile and Link Time 9.94 29.06
Execution Time 5.77 9.51
Object Code Size 274 297
Dhrystone benchmark
Turbo C Microsoft C
Compile Time 5.93 30.70
Compile and Link Time 10.60 44.22
Execution Speed(Dhrystones/second) 1282 1240
Object Code Size 1164 1375
Benchmarks run on a 6 mhz IBM AT using Turbo C version 1.0 and
the Turbo Linker version 1.0; Microsoft C version 4.0 and the MS
overlay linker version 3.51.
- END -
Mark L. Williams
(mlw@ncsc.arpa)
------------------------------
Date: Mon 23 Feb 87 12:08:19-PST
From: Jackie <Burhans@USC-ECLB.ARPA>
Subject: DisplayWrite3 and ASCII files
Yes, DisplayWrite3 can read ASCII files. Just use the GET command which
you call up by pressing f6 then typing G or moving the pointer to the GET
option on the menu. Then type in the name of the ASCII file and it will
be imported. When you save, it will save (under a new name) in DW3
format.
DW3 can also produce ASCII files by pressing f4, selecting the ASCII_Save
option from the resulting menu and moving the cursor to the end of the
block of text you want to save.
Very easy stuff.
JACKIE BURHANS
BURHANS%ECLB@USC-ECL.ARPA
------------------------------
Date: Mon, 23 Feb 87 15:31 EST
From: Thomas Kunselman <CCSTK%CCOL.BITNET@wiscvm.wisc.edu>
Subject: DisplayWrite3 and ASCII files
In response to Betty Harvey's request about a public domain program for
converting ASCII files to DisplayWrite 3 format.
This is built into DisplayWrite 3. Just use the F4 (block key) and
use the RECALL option to recall a block of text previously saved as
either a DisplayWrite 3 document or as a standard ASCII file.
(see page 3-21 of the DisplayWrite 3 User's Guide, Volume 1)
jon
(Workaholic, 98% Caffeine / 2% Human)
------------------------------
Date: Mon, 23 Feb 87 15:40 EST
From: UEG101%URIMVS.BITNET@wiscvm.wisc.edu
Subject: PC To Apple Disk transfer
Hello,
A Few issues back, somebody inquired about a product that will
transfer files from an Apple IIe to an IBM-PC. Apparently, such a
board exists for the I from PC Week's "Ask PC-Week" 2/17/87, Vol 4. #
7, p126
ASKY inc (430 stevens Creek Blvd, suite 287, San Jose Calif, 95129,
(408)274-5724) manufactures the Envoy Disk Controler , an add-in card
that lets the apple read and write to PC MS-DOS disks. The Dis
retails for $180.00
Hope this can be of some help.
------------------------------
Date: 24 Feb 1987 15:15:28 PST
Subject: Program to Edit Binary Files
From: Billy <BRACKENRIDGE@C.ISI.EDU>
To: nomdenet@VENERA.ISI.EDU
Any debugger will do this including debug that comes with DOS. Also the
Peter Norton Utilities do this. Who knows there might even be an adb for
DOS floating around perhaps someone will know of one.
------------------------------
Date: 23 Feb 87 13:51:46 PST (Monday)
From: Bicer.ES@Xerox.COM
Subject: MSC Floating Point Bug?
I noticed something funny under MSC 3.0 and 4.0
int j, k;
...
...
k = 800;
...
j = k * 3.1;
OR
j = ((double) k) * 3.1;
OR
j = (int) (((double) k) * 3.1);
printf ("%d\n", j);
j is 247 instead of 248. A bug in MSC?
Jack Bicer
Bicer.ES@Xerox.COM
------------------------------
Date: Mon, 23 Feb 87 14:56:04 PST
From: Bruce_A._Cowan%SFU.Mailnet@MIT-MULTICS.ARPA
Subject: Bug in Microsoft MASM or DOS 3.2 or RAMdisk drivers?
I have a friend who recently reformatted his 40 meg hard disk and set it
up as two partitions with EVDISK from Everex. Everything seemed ok, but
operations to the root directories of both partitions would screw up in
various interesting ways. He resolved the problem by rebuilding the
hard disk again from scratch. So, something went wrong somewhere in the
build process.
I have an IBM PC-AT with a PRIAM 60 meg hard disk and I use the PRIAM
EDISK driver to do the same thing. I also have similar problems on my D:
disk, but they are very few and far between and I am waiting for a new
release of the EDISK driver and associated software. On the phone PRIAM
expressed amazement that I didn't have worse problems with the old version
of the software I had.
------------------------------
Date: 24 Feb 87 09:27:00 GMT
From: nortond@well.UUCP (Daniel A. Norton)
Subject: SWPTR.COM Swaps LPT1 & LPT2
I keep seeing a mention of swapping printer ports LPT1 and LPT2. On
my PC, the following program works just fine. Copy the file between
the "CUT HERE" lines to a file named swptr.deb and execute the
command:
debug <swptr.deb
This will create the file swptr.com which swaps LPT1 and LPT2. Run
the program twice and it will flip them back to there original
places. If you have spooler software, you should execute this
command before loading it or before the first time you execute the
DOS "print" command.
------------------------------ CUT HERE ------------------------------
a100
mov ax,40
mov es,ax
mov bx,8
es:
mov ax,[bx]
es:
xchg ax,[bx+2]
es:
mov [bx],ax
int 20
rcx
0014
nswptr.com
w
q
------------------------------ CUT HERE ------------------------------
Daniel A. Norton ...!lll-lcc!{lll-crg,ptsfa}!well!nortond
------------------------------
Date: 24 Feb 87 09:27:09 GMT
From: nortond@well.UUCP (Daniel A. Norton)
Subject: MSC 4.0 "system" function
The MSC "system" function does not work quite as one would hope it
would. In particular, it never returns the exit code (a.k.a.
errorlevel) from the called program, but returns the exit code from
COMMAND.COM instead. This value almost always indicates success
unless you try to pass it too many characters on the command line.
I have found the following to work better:
1) strip the program name from the command line and call
"spawnlp" instead of "system"
3) If spawnlp returns anything but -1, the program executed
and the returned value is the errorlevel.
3) if spawnlp returns a -1, it is because it could not
find the named program. Since the command could be an
internal command (such as dir), NOW use the system call to
execute the command.
This does not cover the case when the command is pure garbage.
The program will not be found and "system" will not return an error
in response to the garbage. But at least the above approach is a
step in the right direction.
Daniel A. Norton ...!lll-lcc!{lll-crg,ptsfa}!well!nortond
------------------------------
Date: 24 Feb 87 09:27:15 GMT
From: nortond@well.UUCP (Daniel A. Norton)
Subject: LA-50 on the Serial Port
Don't forget to insure that DSR and CTS are asserted on the
RS-232 connector when the system boots (I tie these lines to DTR).
Otherwise DOS will not look at the serial port. ProComm, however,
doesn't generally care about these signals, which explains why your
printer works with it and not with DOS.
Daniel A. Norton ...!lll-lcc!{lll-crg,ptsfa}!well!nortond
------------------------------
Date: Tue, 24 Feb 87 22:58:46 EST
From: jcm@ORNL-MSR.ARPA (James A. Mullens)
Subject: Another Bug in Microsoft MASM 4.0
RE: Volume 6, Issue 10, From: Ya'akov_Miles%UBC.MAILNET@MIT-MULTICS.ARPA
>Subject: Another Bug in Microsoft MASM 4.0
>There appears to be a bug in Microsoft MASM assembler ver 4.0, as
>is demonstrated. Note that the ASSUME directive seems to require
>that the segments be specified in a specific order, or forward
>references from CALL/JMP do not work properly...
>
> Title This does not work
> ASSUME CS:myseg,DS:myseg,ES:myseg,SS:myseg
>myseg SEGMENT
> call abcd
>abcd: ret
>myseg ENDS
> END
>
> Title This works
> ASSUME DS:myseg,SS:myseg,CS:myseg,ES:myseg
>myseg SEGMENT
> call abcd
>abcd: ret
>myseg ENDS
> END
I cannot explain MASM's choice of error message ("near call to
routine in different segment" -- obviously a wrong call). However,
my assembler manual says the ASSUME directive must name a segment
*previously defined* with the SEGMENT directive. In this case,
ASSUME names "myseg" before it is defined. If I move the ASSUME
after the SEGMENT declaration, the module assembles without error.
If this is the "problem", then the error message should be "reference
to an undefined segment" at the ASSUME statement.
MASM does let you get away with this on the DS, ES, and SS registers,
as far as I can tell.
As to why the second test case worked? ... I guess you should feel
grateful about that! ;-)
jim mullens / jcm@ornl-msr.arpa / oak ridge national laboratory
------------------------------
Date: Mon, 23 Feb 87 17:16:53 est
From: Paul Fishwick <fishwick%ufl.csnet@RELAY.CS.NET>
To: info-ibmpc@C.ISI.EDU
Subject: Diagnostics
I am using DIAGS and have noted that a chip on my system board must be
bad (I am getting an ERROR 131). Do you have a list of these numbers and
the corresponding chip locations? I think I saw something on the digest
a while back (and even thought that I had saved it!) --- could you
possibly ship me a copy of the error list? Thanks much. It is an old
64KB motherboard in case that makes any difference in codes...
-pf
[The list is somewhere but I don't think it covers 64K motherboards. -wab]
------------------------------
Date: Tue 24 Feb 87 07:49:31-PST
From: Steve Dennett <DENNETT@SRI-NIC.ARPA>
Subject: KISS Laser Software
There are a number of software programs around to easily set options
(fonts, etc.) for the HP laser printer. Does anyone know of one that
does the same for the KISS laser printer? Thanks.
Steve Dennett
dennett@sri-nic.arpa
------------------------------
Date: Tue 24 Feb 87 12:27:32-PST
From: Steve Dennett <DENNETT@SRI-NIC.ARPA>
Subject: SWPKEYAT.ASM
I recently ftp'd the SWPKEYAT.ASM from SIMTEL-20. It is a program that
is supposed to reverse the [~] and [Esc] keys on the AT keyboard.
However, after having a friend assemble it and running it, the only
effect it has is to disable the keyboard completely. Same result
occurred on both an IBM AT (older, 6mhz model) and an AT clone.
Has anyone gotten this program to work successfully? If so, maybe
they could put an assembled copy somewhere FTPable? Thanks.
Steve Dennett
dennett@sri-nic.arpa
------------------------------
Date: Tue, 24 Feb 87 15:35:17 est
From: catone@dsl.cis.upenn.edu (Tony Catone)
Subject: Copying 0 length files
Andrew Malton <malton%csri.toronto.edu@RELAY.CS.NET> writes:
>the COPY command seems to refuse to copy empty files. Is this right?
>Is it a bug? Is there a way to do it? Can somone mail me an explanation
>please?
The copy command does refuse to copy files of 0 length. This was used by some
companies very early on as a primitive form of copy protection. It is easily
defeated once you realize that the diskcopy command makes an exact duplicate
of the source diskette, and thus copies 0 length files along with everything
else (even "unused" sectors, useful when you've accidentally erased a disk and
need to try several different recovery schemes).
- Tony
catone@wharton.upenn.edu
catone@dsl.cis.upenn.edu
------------------------------
Date: Tue, 24 Feb 87 15:36:38 est
From: catone@dsl.cis.upenn.edu (Tony Catone)
Subject: Disk optimizers
The info-pc lending library has a disk optimizing program written in Turbo
Pascal (with lots of in line code) called Reformat, available via anonymous
FTP. I have used it on a variety of PC's and AT's, always without a glitch.
It does, however, require you to uninstall copy protected software like Lotus
123. This usually isn't a big problem, though, as disk optimization is not an
everyday affair.
- Tony
catone@wharton.upenn.edu
catone@dsl.cis.upenn.edu
------------------------------
Date: Tue, 24 Feb 87 16:38:35 cst
From: cody@anl-mcs.ARPA (Jim Cody)
Subject: Wierd problem with Mace Utilities disk
I have just purchased a copy of Mace Utilities for use on
my Leading Edge Model D, and have encountered a wierd problem.
Booting from the C drive, inserting the Mace Utilities disk in
the A drive, and typing DIR A: results in the standard "Drive
not ready" error. The disk has been tried on a second Model D
(at a dealers) with the same results, but causes no problems
on any other non-Leading Edge machine I have tried. The disk
is fine in all other respects; i.e., a copy of the disk made on
another machine works perfectly on my Model D.
That behavior rings a bell; somewhere I have seen something about
certain types of disk drives malfunctioning on disks with "shiny
envelopes," but I have not located the source of that information
with a quick search of recent PC-Digests.
I am otherwise satisfied with the product. The usual disclaimers
apply.
------------------------------
Date: Tue, 24 Feb 87 17:41:11 EST
From: dewar@nyu-acf2.arpa
Subject: Buffers
I don't know how thrasher works, but I have an interesting recent experience
which shows how delicate the number of buffers is.
To save space on my DP386, I decided to cut down the number of buffers from
50 to 20, I thought 20 should still be enough for reasonable performance.
Well the performance for loading files got dramatically worse. With a little
more experimentation, I found a huge gap between 24 and 25 buffers. What was
happening is that I have a huge directory full of all my executables. With
less than 25 buffers, this directory did not fit in memory, so any search
for a program reasonably far on in the directory was a worst case search
(LRU is a pessimization algorithm if you are doing serial searches over
a file which does not fit into available memory).
This example shows how delicate the choice of number of buffers can be. I
have NEVER seen a case where increasing the number of buffers measurably
decreases the performance of a system (despite the warning in the DOS
manual). On the contrary, I often see that 99 buffers gives measurably
better performance than 60 buffers. So I think the general advice on
buffers is simple: use the biggest value for buffers that you can, given
your other memory requirements!
------------------------------
From: zhahai%gaia.UUX%ncar.csnet@RELAY.CS.NET
Date: 24 Feb 87 10:21:19 GMT
Subject: Use 2 8250's to get 1200/75
I rather doubt that the 8250 UART chip will handle being switched between
the 1200 and 75 bps rates for split speed modems. If you can guarantee
half duplex without reverse channel signalling this could work (re-init
before each direction change), but for full duplex you would ruin the
characters coming in at one rate when you reset the speed for the outgoing
rate. As wab says, a buffered modem with always speaks 1200 to the PC
would be the most elegant solution. Failing that, many PC's have or can
have two serial ports - nothing says you can't wire up a split cable to
connect recieve from the modem to one, xmit to the other - and program each
8250 to a different rate. Again, not a universal solution, but it might be
good for some people.
By the way, the 8250 as used by IBM for serial communications does not have
to run at the rates given in the IBM manuals. It will run at any bps found
from 115,200/n, where n is 1..65535. If you want 1234 bps, just do the
divide and round - the speed may not be exact, but it usually will be within
the range needed for async unless the speed is very high (high rates = small
divisors = large granularity; 9600 divisor is 12, so you can only adjust it
be +/- 8% or so). ~z~
Zhahai Stewart
{hao | nbires}!gaia!zhahai
------------------------------
From: ihnp4!gargoyle!sphinx!goer@ames.arpa (Richard L. Goerwitz III)
Date: 24 Feb 87 05:08:55 GMT
Subject: Herc Plus SIMCGA Compatibility
Organization: U. Chicago Computation Center
Has anyone else experienced problems running SIMCGA on the new Hercules Plus
video card? It's a wonderful device, allowing twelve different fonts in text
mode (fast). Included are drivers for some of the more popular PC/XT/AT pro-
grams, such as Microsoft Word (which resorts to bit-mapped (slow) dis-
play so as to do, say, italics). Trouble is that as soon as I ran SIMCGA
on it, it crashed. In other words, my old Hercules-compatible graphics card
used to run SIMCGA. Now my supposedly downward-compatible H Plus won't run
it.... Now before I go to great lengths determining just what is wrong, I'd
like to see if I can get some input from others who have Hercules Pluses.
It might be malfunctioning hardware. More probably, however, SIMCGA is going
to have to be modified. But in either case, I'd like to know first whether
others are experiencing the same problem.
TIA (Thanks in advance),
Richard L. Goerwitz
P.S. I prefer replies via
bitnet (goer@sphinx.uchicago)
but !ihnp4!gargoyle!sphinx!goer
is okay....
------------------------------
Date: Wed, 25 Feb 1987 10:43 PST
From: JAJZ801%CALSTATE.BITNET@wiscvm.wisc.edu
Subject: Btrieve
I have been using Btrieve for about a year now. You should
be aware that it is NOT really a database product. It is
a file manager. That it, it recognizes data only at the
record level as a defineable unit, not individual fields.
Therefore, it in no way could read or manipulate lotus 123
files. Moreover, it only reads external ascii files (through
a utility program) for loading files.
It does have many useful feature such as caching of records
logging of transactions to ensure file integrity in the event
of a system crash. It is, however, designed not as a stand-
alone product but to be incorporated with programs, on a called
basis. There are interfaces for MANY languages.
There are two add-on products from the manufacturer, called
Xtrieve and Rtrieve which permit field-level definition and
manipulation. The letest database review of PC magazine
(vol 5, no. 12, June 24, 1986, p.190) listed it as a relational
system (with the add-ons) but did not exactly give a glowing
endorsement. I have not heard of any ability to read lotus
files for the add-ons but the company does not send out a
lot of promotional material. I will check their bulletin
board about it and send you another message.
Overall, I am satisfied with Btrieve but my purpose and use
was to mimic a min-computer's file management environment with
minimum changes to programs (I modifed the file-manager-specific
routines that interfaced to the O/S and was able to leave the
main programs largely unchanged).
You do not mention what your overall goal or need is, but
there are a number of database systems (not all of which I
know of) which can import lotus data, including spreadsheet
clones of lotus which can access dBase files.
Jeffrey Sicherman
Calif. State Univ., Long Beach
JAJZ801@CALSTATE.BITNET
------------------------------
Date: Wed, 25 Feb 1987 10:50 PST
From: JAJZ801%CALSTATE.BITNET@wiscvm.wisc.edu
Subject: MS Fortran Version 4.0
It may be a little early to ask this question since I just myself
received the update offer for MS Fortran 4.0 (from 3.31), but maybe
there are beta users out there.
Has anyone had any experience with this version yet. I'm mainly
interested in what the real advantages of upgrading are (aside from
the propoganda claims of official ANSI and govt certified (as if that
were a guarantee of anything anymore). What do a *gain*, not get, from
upgrading aside from the negative gain from my bank account to Bill
Gates' of $150. Also, what if any changes might be necessary to
programs that compile and run under 3.31 to work with 4.0. Will
it be necessary to compile any/all subroutines ?
Thanks for any advice.
Jeff Sicherman
JAJZ801@CALSTATE.BITNET
------------------------------
Date: Wed 25 Feb 87 14:52:35-PST
From: PAWKA@NOSC-TECR.ARPA
Subject: MS-DOS Using Extended Memory
Now that I've installed 2560K extended memory in my Z-248,
is there any way to use it other than RAM disk? Apparently MS-DOS uses
the upper portion of memory for the transient portion of COMMAND.COM. Is
there a patch to move these addresses up into extended memory?
Mike Pawka
PAWKA@NOSC-TECR.ARPA
------------------------------
Date: 23 Feb 87 10:50:00 EST
From: "MAGDI N. KAMEL" <kamel@wharton-10>
Subject: ARC51.COM DOWNLOAD PROBLEMS
I have been trying to download the archival program ARC51.COM from
SIMTEL20. I successfully FTP'ed the program to our VAX 8650 using
binary transfer mode. I was also successful in downloading the
program from the VAX to my IBM AT using Kermit. When I tried to run
the program I got 'Program too big to fit in memory' message. Does
anybody know what is the problem. My arpa address is:
KAMEL@WHARTON.ARPA Thanks.
------------------------------
Date: 23 Feb 87 07:56:42 PST (Monday)
Subject: MINIX dl
From: Hodges.osbunorth@Xerox.COM
Does anyone know if arpanet/usenet dl devoted to MINIX is an arpanet dl
or a usenet dl? Anyone know what the complete name of the dl is?
Andy Tanenbaum indicated the dl is named "comp.os.minix" in one of his
msgs in Info-IBMPC Digest, but that name is incomplete from our view of
the world.
Any help appreciated.
jeff
[comp.os.minix as a name has meaning only if you are on a unix machine
on usenet running the rn program. -wab]
------------------------------
Date: 24 Feb 1987 15:08:05 PST
Subject: MINIX dl
From: Billy <BRACKENRIDGE@C.ISI.EDU>
To: Hodges.osbunorth@XEROX.COM
We have tried repeatedly to get someone to take responsibility for internet
distribution of this discussion group but there are no volunteers.
------------------------------
Date: Mon 23 Feb 87 11:21:49-PST
From: Bob Knight <KNIGHT@SRI-NIC.ARPA>
Subject: Reentrant ROMable 8087 Emulator Supplier?
Hi all - I am looking for an 8087 emulator that is both reentrant and
ROMable, and useable with the Intel compilers (PLM86 and ASM86 to be
specific.) As far as I know, Intel's library is non-reentrant, and,
even more disgustingly, uses initialized data.
Any and all information (as well as [polite] corrections if I am
misinformed regarding the Intel emulation) would be very much
appreciated.
E-mail directly to me, please don't burden the lists. I will summarize
if there's interest.
Thanks,
Bob
KNIGHT@SRI-NIC.ARPA
------------------------------
Date: 23 Feb 1987 14:00:53 CST
From: CCSO-ALL@GUNTER-ADAM.ARPA
Subject: PMAP VER. 1.1
I would appreciate any information that anyone can give me a program
witten in 1986 by C. J. Dunford called PMAP. Any information on where
I can find the program or who I can contact to get will much
appreciated.
------------------------------
Date: Mon, 23 Feb 87 14:15:54 EST
From: Jim.Moore@WB1.CS.CMU.EDU
Subject: Clock Calendars
Do you have a reccomendation for a clock/calendar for someone who already has
all the memory they want, needs no additional printer ports or joysticks?
Either a c/c with nothing else (cheap!) or a c/c with some specifically
useful (other) function(s) (also cheap!)?
Thanks,
Jim
------------------------------
Date: Mon, 23 Feb 1987 13:15 PST
From: JAJZ801%CALSTATE.BITNET@wiscvm.wisc.edu
Subject: From Addresses
Occasionally, From addresses appear without a network domain in them
(the .xxxx part, forgive my terminology, I'm not a networks expert). I
realize there's propably no way to force the inclusion though it would
be useful for personal replys. Is there some assumption that can be
made of the network component that appear without one based upon the
routing to you ?
Jeff Sicherman
JAJZ801@CALSTATE.BITNET
[ Life would be easier on all of us if people would get with the program
and implement the domain system on networks all over the world. Just for
reference .BITNET is not a valid domain. ISO country codes, .COM, .EDU,
.ARPA, .UUCP are valid top level domains. Domain names do not refer to
networks they refer to organizations. When I answer your mail my system
shouldn't have to know what network you are on. All I want to know is the
name of an organization that knows how to deliver mail to you.
Please folks pester your local system administrators and network
political types to join the party or get left out in the cold. -wab]
------------------------------
Date: Fri, 20 Feb 87 20:12:55 GMT
From: A385%EMDUCM11.BITNET@wiscvm.wisc.edu
Subject: Amstrad PC-1512
HELLO NET!!
I would like to buy an Amstrad PC-1512 but I've heard that there are a lot
of problems with the hard disk.
Does anyone know anything about this or other problems with the PC-1512?.
Thank you very much in advance.
Javier Lopez Torres
Universidad Complutense de Madrid (Facultad de Matematicas)
28004 Madrid (Spain)
<A385@EMDUCM11>
------------------------------
Date: Mon, 23 Feb 87 22:08:49 EST
From: Chris Schmandt <geek@MEDIA-LAB.MEDIA.MIT.EDU>
Subject: EMM for PC-elevATor
Has anyone had any success with the PC-elevATor accelerator card's
expanded memory manager? I'm trying to run Golden-Bow's VCACHE,
which is otherwise reputable, on it and boy, do I lose quick!
thanks
chris
------------------------------
Subject: Program to Edit Binary Files
Date: Mon, 23 Feb 87 10:31:07 PST
From: nomdenet@venera.isi.edu
Is there a program which allows modifying binary files on PCs?
I want something like Unix's adb.
Thanks,
Bert White
------------------------------
End of Info-IBMPC Digest
************************
-------