[comp.sys.ibm.pc.digest] Info-IBMPC Digest V91 #82

Info-IBMPC@WSMR-SIMTEL20.ARMY.MIL ("Info-IBMPC Digest") (04/10/91)

Info-IBMPC Digest           Thu,  4 Apr 91       Volume 91 : Issue  82 

Today's Editor:
         Gregory Hicks - Rota Spain <GHICKS@WSMR-Simtel20.Army.Mil>

Today's Topics:

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: Thu, 28 Mar 91 11:16 EST
From: Michael Gordon <GORDONM@MAX.CC.DENISON.EDU>
Subject: Borland (a comment)

>In addition, Paradox is from Borland.  It is a stable product, from a
>stable company so you don't need to worry about the future of your
>investment.  Hope this helps.

Would users of Sprint agree? 

Borland produces fine products (I'm a satisfied user of Quattro,
Sidekick, Paradox, and Reflex), but this doesn't necessarily mean that
all of their products are necessarily good investments for the future.
All indications are that Reflex is going to go the way of Sprint and
that Borland is only going to concentrate on Quattro/Paradox and, given
the recent upgrade, Sidekick. (I hope I'm wrong regarding Reflex.)

Michael D. Gordon               		Internet: gordonm@cc.denison.edu
Department of History                           Bitnet:          gordonm@denison
Denison University                              VoiceMail:        (614) 587-6588
Granville, OH 43023                             FAX:              (614) 587-6417

------------------------------

Date: Fri, 29 Mar 91 08:28:41 PST
From: stewart%watop.nosc.mil@nosc.mil (Stephen E. Stewart)
Subject: Bugs in MS FORTRAN V5.0 using 8087

In comp.sys.ibm.pc.digest you write:

>In Reply to this Note From: <C102TA%AINUNI01@pucc.PRINCETON.EDU>
>>Like Joe Landman, I also found a strange behaviour of MS FORTRAN 5.0
>>working under 8086/8087 machines. [text deleted]

>>Has anybody made similar observations?  Does anybody know what to do
>>in this case?

>It is not uncommon for bugs to hide when the debugger is loaded.  It
>[text deleted]

I just noticed this thread so I missed the original problem.  I too
have had some difficulty with MS FORTRAN V5.0 using a 80387.  If I
linked with the 80387 hardware math library the program would not run,
but if I linked with the emulation library (which would use the
hardware if present) it would run.  I do actually have a functioning
80387 in my machine.  Do you know enough about Joe Landman's problem to
say whether my problem might be related?  Who else could I email to
discuss this?

Steve Stewart                                    Naval Ocean Systems Center
stewart@watop.nosc.mil                           San Diego, California

------------------------------

Date: Fri, 29 Mar 91 11:28:14 EST
From: 07480JLN%ibm-fal.cl.msu.edu@msu.edu
Subject: Bugs in MS FORTRAN V5.0 using 8087

    I have recently received from MicroSoft the application note
7/90-HF0327, detailing a BUG  in their handling of the 8087 control
word.  This application note came with a disk.  I followed the
directions on the disk, and lo-and-behold... My rootfinder found roots,
and the savage benchmark worked, from DOS.

        My original problem never was a bug in my code, but rather an
interesting behaviour of programs on 8088/8087 based machines.  In the
original post, I mentioned that the program worked properly on all
other machines, and on the 8088/8087 based machines under a different
compiler.  Thus I correctly reasoned that the program was functioning,
but the compiler somehow wasnt working properly.   I investigated its
behaviour, and it appeared to me that the problem was in dealing with
roundoff.  That is, I found an accumulation of roundoff error (as in
the savage benchmark), that was sufficient to swamp my code.  After I
thought I understood the bugs behaviour, I then posted to the rest of
the networld and asked if anyone else had observed it.  I received
about 10 responses, ranging from "yes, I have seen it", to "no, you
must be a bad programmer..".

       Well, I know that the situation did exist, and the patch
corrected it.  Microsoft has stated that they allow one to distribute
the patch "with all documentation".  If they simply allowed a zip-file
to be posted to simtel, I would do that.  Since it comes with this
application note, I suggest the following.  Contact MicroSoft at
(206)-454-2030, and ask them for a copy of this application note I
mentioned, regarding a patch for the 8087 control word setup library
module.  I can now say that I trust the Microsoft compiler as far as
anyone can trust a compiler, to be free of bugs that will effect my
work.  Only if they had a QuickFortran not unlike their QuickBasic...

      Joe Landman  Graduate Student: Computational Condensed Matter Physics
                   Wayne State University dept of Physics.

        Internet:     userzv50@mts.cc.wayne.edu
                      07480JLN@CMS.cc.MSU.edu
        Bitnet:       userzv50@waynemts.bitnet
                      07480jln@MSU.Bitnet
        ATTnet:       (313)-577-2752
        snailnet:     666 Hancock Avenue, Detroit MI 48202

------------------------------

Date: Fri, 29 Mar 91 10:20:45 EST
From: Scott Begin <34RXESC%CMUVM.BITNET@CUNYVM.CUNY.EDU>
Subject: dBase and Epson Printers: (Vol.91 No.69)

In Vol. 61 #69 Walt <lazear@gateway.mitre.com> writes about problems he is
having with an Epson LQ 510 printer using dBase.

The problem is not your printer, it is a not real well documented
limitation of dBase:  dBase will not send a null {CHR (0)} chraracter
to a printer.  It is mentioned in the dBase III manual where this
function is described, but not real clearly in my dBase III plus
Programmer's Reference Guide (by Sybex).

If you can't send a null character to a printer, how do you access
those printers functions called by a null character?  I have found a
couple of solutions:

In the dBase III manual describing the CHR function, it mentioned a
technique for turning off the high bit of the character and sending a
CHR(128).  I don't have my manual handy and I havn't tried this
technique.

The second method comes from a booklet that came with my Epson LQ 1010
printer called "Application Notes  Using Espon Dot Matrix Printers with
{list of software ommitted}"  It has a price of $6.95 on the cover, as
well as a number Y90299107001.  I have found it very useful.  Ask your
Epson Dealer about getting one. [Standard Disclaimers Apply]

It claims that instead of sending a null CHR(0), that you can send a 0
CHR(48) for some commands, althought some commands will not allow this
(see later description).

The third method (mentioned in the booklet) calls for creating a binary
program file, loading it during your application and calling that when
you want to send a null character to the printer.  The binary program
contains the following characters (numbers are decimal):

184 000 000 186 000 000 205 023 203

I created this file using a BASIC program, although it could be created
using DEBUG, or something similar.

When I loaded this file into DEBUG and unassembled it, it came up with:

MOV    AX,0000
MOV    DX,0000
INT    17

To use, it recommended something like:

LOAD NULL
SET PRINT ON
SET CONSOLE OFF
? chr(27)+chr(45)+chr(1) && Turn Underline printing ON
?? "Underline On"
? chr(27)+chr(45)
CALL NULL            && Turn Underline printing OFF
?? "Underline Off"
SET PRINT OFF
SET CONSOLE ON

The last method is the one I use.  It takes advantage of the Epson
Master Select function (ESC !).  This is one place where when you need
a null, you must send a null character, not a 0 character.

The Master Select controls 9 other mode toggles with a single command.
The use is ESC ! and a number telling which modes you wish to turn on
or off.  To calculate the number, add up the values for the following
functions which you want to turn on:

Pica (10 Pitch)         0
Elite (12 Pitch)        1
Proportional Spacing    2
Condensed print         4
Emphasized Print        8
Double Strike          16
Double Wide            32
Italic Characters      64
Underlining           128

If you want underlined elite type (and all others turned off), the
command is:

? CHR(27)+CHR(33)+CHR(128+1)

To return to elite non underlined type, the command is:

? CHR(27)+CHR(33)+CHR(1)

This still presents problems when you want to turn pica 10 pitch on and
the rest off (requires a ESC ! chr(0)), but I have  gotten around this
by issuing the above command for elite type followed immediately by the
command to select pica type:

?? CHR(27)+CHR(80)

One word of caution about using the Epson Master Select, it does not
work for all Epson printers.  I have found it to work on my LQ 1010, an
LQ 800, an FX 286e (and a Sperry printer emulating this printer), but
it does not work on an Epson MX 100-III, and I don't remember if it
worked on an Epson FX 100 and FX 100+.  I expect it would work on all
LQ Series printers, all the current Epson dot matrix printers and
anything introduced in the past several years.

I hope this helps...

   Scott A. Begin             E-Mail:      34RXESC@CMUVM.BITNET  (Network     )
"The Graveyard Madman"                     CSV310                (CMU VAX 8530)
  Student Consultant          AT&T:        (517) 773-9424
  Ronan Computer Lab          USnail:      1401 E. Bellows #E-1
Central Michigan University                Mt. Pleasant, MI  48858

------------------------------

Date: 28 Mar 91 22:16:05 EST
From: James O'Donnell <JODONNEL@PENNSAS.UPENN.EDU>
Subject: Info-IBMPC Digest V91 #68

Sorry to be obtuse, but I'd be glad to hear this in so many words. If I
run KERMIT under WINDOWS or DesqVIEW, may I leave KERMIT uploading or
downloading and go cheerfully and multitaskingly off to do my other,
e.g., word processing chores? I've been unconvincingly advised that
this is not possible and would be glad to hear otherwise. Any cautions
or problems?

J.J. O'Donnell
JODONNEL@PENNSAS.UPENN.EDU
University of Pennsylvania

------------------------------

Date: Fri Mar 29 10:40:00 1991
From: johnboyd@logdis1.oc.aflc.af.mil (John Boyd;CRENP)
Subject: KA9Q, etc...

What is the KA9Q pkg&utils that I see discussed occasionally.  Looks
sort of like a 'HAM' call sign. Is it packet software?  I know,
curiosity killed the cat B-).

[The KA9Q TCP/IP Package is a complete implementation of the TCP/IP
protocols for MSDOS.  Capabilities include TELNET, FTP, PING, SMTP,
...

KA9Q is a Radio Amateur Call Sign.  gph]

------------------------------

Date: Fri, 29 Mar 91 11:30:53 EST
From: George Waller <HBLADM1%UCONNVM.BITNET@YALEVM.YCC.Yale.Edu>
Subject: Another floppy drive problem

What might cause this on a Zenith 248 (80286 IBM clone)?

At the C prompt I issued the command MORE<AUTOEXEC.BAT and the result
was the A disk drive light came on and I got the Abort, Retry, Fail
message. There is no reference to the A drive in PATH.

George Waller, Microcomputer Consultant           HBLADM1@UCONNVM.BITNET
Homer Babbidge Library, University of Connecticut,  Storrs CT 06269-1005

------------------------------

Date: Thu, 28 Mar 91 15:18 -0600
From: DARIN NEVEN <N3CMW@CCVAX.IASTATE.EDU>
Subject: C Books

I need to learn C for a summer engineering job.  I looking for a good
book that introduces and teaches the language.  I know have programed
in a few other languages.  I would appreciate any recommendations on a
good book about C.
 
Thanks in advance

n3cmw@ccvax.iastae.edu
Darin Neven
Iowa State University
Chemical Engineering 

------------------------------

Date: Fri, 29 Mar 91 08:54:37
From: BS04@primeb.dundee.ac.uk
Subject: Difference between Expanded and Extended Memory

Networkers,

 I have recently bought a British IBM clone (Viglen 286) and I am
really quite happy with it. One thing I would like some help on is the
difference between extended and expanded memory. Hope this question is
not too close to the dum level. How do you check the set up of the
memory and can anyone recommend a book on the subject?

Yours
Andy Johnston.

Andrew M. Johnston,                    E-Mail: BS04@PRIMEB.DUNDEE.AC.UK
Biological Sciences,                   Tel No: 0382-23181 Ext 4282
University of Dundee,
Dundee
Scotland
DD1 4HN

------------------------------

Date: Fri, 29 Mar 91 13:13:36 SET
From: Alexander Falk <K360950%EDVZ.UNI-Linz.AC.AT@pucc.PRINCETON.EDU>
Subject: Number Crunching - Help needed

My department is currently planning to buy a machine for simulation of
x-ray diffraction curves. What we need is real number crunching power
(about 10MFlops minimum) and we have a budget of about $20000.

What we'd like to do is buy either some DOS PC (486) or a Macintosh
IIfx and install an add-on board for the number crunshing jobs (i.e. an
i860 or any other RISC chip, trnasputer, etc.). Does anyone have
experience with such add-on boards? Should we rather invest our money
in a "true" workstation (some Unix machine from Sun, HP, or even the
RISC/6000 from IBM).

Our requirments are: Fortran and C compiler for the number crunching
chip, which allows one to run code at about 10 MFlops (or better). If
the machine can be used for other tasks it would be nice, but is not
necessary.

Any help is appreciated,

Alexander Falk

------------------------------

Date: Sat, 09 Mar 91 17:18:48 -0500
From: (Marshall D. Abrams) <abrams@smiley.mitre.org>
Subject: Send-FAX software

A fried has purchased a Incomm 24FXE modem that can send (but not
receive FAXes.  That is, the modem is CCITT T.30/T.40 compatible.  It
came with software for the PC known as BIT-CHAT, version 2.0

BIT-CHAT appears minimally functional.  It doesn't display the
acknowledgement information from the remote FAX.  It appears to have a
problem with dashes (hyphen, minus-sign) in the phone number following
the area code.

Can anyone recommend other software, preferably on Simtel, that would
work with the modem but would provide more functionality?

Sincerely,
 
- Marshall D. Abrams, phone: (703) 883-6938
   The MITRE Corporation, 7525 Colshire Drive
   Mail Stop Z269, Mc Lean, VA   22102
   alternate e-mail address: abrams@mitre.org
   FAX: (703) 883-1397  

------------------------------

Date: Fri, 29 Mar 91 11:32:57 EST
From: George Waller <HBLADM1%UCONNVM.BITNET@YALEVM.YCC.Yale.Edu>
Subject: Software needed

Needed: software which will take a DOS data file and insert LF/CR every
n bytes? A word processor would work fine, but these data files are 4-7
MB long! Thanks.

George Waller, Microcomputer Consultant           HBLADM1@UCONNVM.BITNET
Homer Babbidge Library, University of Connecticut,  Storrs CT 06269-1005

------------------------------

Date: Wed, 27 Mar 91 17:03:44 PST
From: raymond@math.berkeley.edu (Raymond Chen)
Subject: IP-HOOK.ZIP - Information on hooking up a PC to the Internet
Summary: Reposted by Keith Petersen

I have uploaded to SIMTEL20

pd1:<msdos.info>
IP-HOOK.ZIP     Information on hooking up a PC to the Internet

On 25 Jan 1991, Dave Cottingham posted a summary of responses to his
request for information on how to hook up a PC to the Internet.
IP-HOOK.ZIP is a ZIP archive containing his summary, plus copies of all
the responses he received.

Raymond Chen
raymond@math.berkeley.edu

------------------------------

Date: Thu, 28 Mar 91 14:05:51 EST
From: srini@IREQ.HYDRO.QC.CA (K.Srinivasan)
Subject: NALKATI.ZIP - Hindu Calendar. Date, positions of Moon & Sun
Summary: Reposted by Keith Petersen

I have uploaded to SIMTEL20:

pd1:<msdos.calculator>
NALKATI.ZIP     Hindu Calendar. Date, positions of Moon & Sun

The Hindu calendar is based on the celestial position of sun and moon.
There are 12 constellations, which divide the sky into equal parts.
The sun stays in a constellation for a Hindu month. The moon stays in a
'nakshatram' (0.44444 of a constellation) for about a day. The birth
day according to the Hindu Calendar is when the sun is in the same
constellation and the moon is in the same 'nakshatram'. Usually, this
does not fall on the same Calendar date. The death anniversary is
observed on the same phase of the moon, which differs from the
'nakshatram'.  This is followed in India, Sri Lanka, Thailand, Malaysia
and Indonesia.

K. Srinivasan, 1109 Rue Gaboury, St. Bruno, QC. J3V 5X4 Canada.
srini@ireq.hydro.qc.ca

------------------------------

Date: Thu, 28 Mar 1991 13:24 +1000
From: "PETER SUMMERS" <U5533129%ucsvc.ucs.unimelb.edu.au@nosc.mil>
Subject: WISDOM.ZIP - A fortune cookie program with Turbo C source
Summary: Reposted by Keith Petersen

I have uploaded to SIMTEL20:

pd1:<msdos.sysutl>
WISDOM.ZIP      A fortune cookie program with Turbo C source

Peter Summers
U5533129@ucsvc.ucs.unimelb.edu.au

------------------------------

End of Info-IBMPC Digest V91 #82
********************************
-------