Info-Mac-Request@SUMEX-AIM.STANFORD.EDU (The Moderators) (04/05/89)
Info-Mac Digest Tue, 4 Apr 89 Volume 7 : Issue 62
Today's Topics:
[DCGQAL]TOM.CHAVEZ!Re: Info-Mac Digest V7 #58
Answers to memory questions
Apple IIe for the MAC
Backup Help Wanted ...
FKEY programming info
Fullwrite Crashes
HELP! Lightspeed C
Info-Mac Digest V7 #55
Leprechaun Demo -- part 1 of 10
Reversi/Othello Help
those darn CLUT's
TK/Solver?
TK/Solver for the Mac
TOPS repeater use with KINETICS FP
Your Info-Mac Moderators are Lance Nakata, Jon Pugh, and Bill Lipa.
The Info-Mac archives are available (by using FTP, account anonymous, any
password) in the info-mac directory on sumex-aim.stanford.edu [36.44.0.6].
Please send articles and binaries to info-mac@sumex-aim.stanford.edu.
Send administrative mail to info-mac-request@sumex-aim.stanford.edu.
----------------------------------------------------------------------
Date: Mon, 3 Apr 89 10:19:24 PDT
From: "DASnet" <XB.DAS@forsythe.stanford.edu>
Subject: [DCGQAL]TOM.CHAVEZ!Re: Info-Mac Digest V7 #58
Re: Subject: MIDI Manager
In the April issue of Electronic Musician there is a reference to Apple
showing the "MIDI Manager" at the winter NAMM show (p. 12). Does anyone
out there have any information on this??? Is it the long awaited fixes to
the Sound Manager, or is it an entirly different package? When will this
be available to the general public, and where can I get it????
Thanks in advance,
Reed Rector
____________________________________
The Apple MIDI Management Tools provide a new way for an application to access
the MIDI interface. The new MIDI driver provides the interface. An
application (or DA for single finder) called PatchBay provides a graphical
interface for connecting ports and timing between applications and devices.
The best feature about the new MIDI Manager is that multiple MIDI applications
are supported under MultiFinder, with real-time interaction and data sharing.
The development package will be available from APDA (the Apple Programmers and
Developers Association) in early May.
If you have other questions, feel free to send me mail.
Tom Chavez
Product Manager, Apple MIDI Management Tools
Apple Computer, Inc.
Internet: Tom.Chavez@applelink.apple.com
=END=
------------------------------
Date: Fri, 31 Mar 89 22:42 EST
From: Robert W. Kerns <RWK@fuji.ila.dialnet.symbolics.com>
Subject: Answers to memory questions
From: John Salmento <ziggy+@andrew.cmu.edu>
It is possilble to use IBM simms in Macs, but not Mac simms in IBMs. IBM simms
are 9 bit chips. The nineth bit is used for parity checking, and according to
Lee Larson,
"statistical analyses have shown that there is no significant
increase in reliability in using that ninth chip as IBM does. It is only
used once, during the boot-up process. It has as much chance of failure
as any other chip in the system, so 1/9 of the time, the parity chip
is the one that goes. (More parts ==> More failures)"
Well, I don't know what IBM actually does, but it is NOT the purpose of
parity checking to improve the reliability. As Mr. Larson said, it does
indeed slightly degrade reliability.
The reason why real computers have parity bits is to eliminate UNDETECTED
errors. It is usually better for the system to crash than for you to get
a wrong answer because of a memory error.
Better than parity checking is ECC -- Error Correcting Code. This uses
additional bits to not only detect errors, but additionally, to correct
the error if it is no more than one erroneous bit. Double bit errors
are detected, but not corrected. Unlike parity checking, ECC can improve
reliability, especially if single-bit errors are logged when corrected,
and any defective boards replaced at a future service visit.
If memory serves, I believe 4 bits is not enough bits to implement ECC
for a 32-bit word, but 8 bits is more than enough; I think it's about
6 bits, but I don't want to think too hard.
ECC is less expensive (proportinally) for larger word widths.
As you can see, ECC is more expensive than parity checking. But I think
it is very bad that Apple does not at least provide parity checking. It
makes it hard to trust results on the Mac, and it makes it hard to identify
problems. Intermittent software problems can really be bad memory, but
how can you tell?
I'm perfectly willing to pay an extra 20% for my memory to get the extra
reliability of ECC.
------------------------------
Date: Sun, 2 Apr 89 21:28 EDT
From: <J_KAZURA%UNHH.BITNET@forsythe.stanford.edu>
Subject: Apple IIe for the MAC
In Vol7 Issue 59,
Harry Bates asked if there was an assembler for the IIe that works
on a MAC. Well there is, its called " ][ in a MAC " It's made by
COMPUTER:applications, Inc.
12813 Lindley Drive
Raleigh, NC 27614
(919) 846-1411
Hope this helps,
Joe Kazura
Computer Specialist
University Technology Center
University of New Hampshire
(Disclaimer: I didn't do it!)
------------------------------
Date: Tue, 4 Apr 89 08:15 CST
From: SWANGER@ducvax.auburn.edu
Subject: Backup Help Wanted ...
I have a small, non-networked Public Mac Lab (8 SE's with 20mb hard drives). I
would like to set up all of our supported software on one of the Macs exactly
like I wanted it, then I would like to use a backup program to create a
'Baseline' set of diskettes from this Mac's hard disk. I would then use this
baseline set of disks to restore each of the other 7 Macs. In theory, each of
the Macs would be set up exactly the same. If some user trashed the hard disk,
it would be easy to erase the disk and restore the software from the baseline
diskettes.
I tried this with HDBackup, Apple's famous program. It seemed to work, except
some file attributes were ignored. Files that were locked on the original hard
disk were not locked on the restored hard disk. Mactools allows you to
'protect' a file. This prevents users from copying an unauthorized (i.e.
licensed) file from the hard disk (well, sort of... if the user has HDBackup
they can copy the file). Files that are 'protected' on the original hard disk
are not protected on the restored hard disk. At least invisible files were
still invisible on the restored disk. I also tried this with FASTBACK (from
Fifth Generation) and met with the same results.
I want a Backup program that will allow you to backup a hard drive to diskettes
and then restore these diskettes to another hard drive. Afterwards, the second
hard drive should be set up EXACTLY like the original drive, not reasonably
close.
I would appreciate any help with this!
David Swanger
Academic Computing Services
Auburn University, Al 36849
205-844-4813
SWANGER@AUDUCVAX <-- BITNET ADDRESS
SWANGER@AUDUCVAX.AUBURN.EDU <-- INTERNET ADDRESS
------------------------------
Date: Mon, 03 Apr 89 00:33:45 EDT
From: David Ascher <ST501649%BROWNVM.BITNET@forsythe.stanford.edu>
Subject: FKEY programming info
This has probably been talked about before, but ``where can I find info
on programming an FKEY for the Mac''? I couldn't find anything in the
TechNotes. Any other idea?
Thanks for any help
David Ascher
E-mail: ST501649@BROWNVM.BROWN.EDU (ARPANet/BITNet)
SnailMail: P.O. Box 3209, Brown University, Providence RI 02912
NewEnglandTelNet: (401) 863-6603
# include disclaimer.h;
Flames, mail, and love letters gladly accepted.
------------------------------
Date: 3 Apr 89 13:46 EDT
From: Joe_Murphy.CAC.CAC@a.darpa.mil
Subject: Fullwrite Crashes
We have a user here who has had repeated crashes with Fullwrite when she
prints. After several crashes, we reload the system files, the Fullwrite
files and network drivers/printer drivers (anything which the crash
might have corrupted) but the problem inevitably reappears. Today we
swapped out the cpu and replaced it with a new one, but I expect the
crashes to continue. Machine crashes whether under Multifinder or
Finder.
Configuration is:
Mac II 4 megabytes or RAM
SuperMac Video 8 board
CMS 60 (was Dataframe 60)
Local Talk connection to 3Com network
System 6.02, LaserWriter 5.02, 3Com 1.3.1
Dimmer INIT, SuperClock, Responder and init CDEV.
FullWrite 1.0
I suspect Fullwrite, but several others are using it without problems.
The machine crashes with various error codes, ID=3 to ID=11. This only
happens when she prints. We also have had the 3com mail program crash
when the machine prints. I suspect something about the software
configuration is causing the problem. Does Fullwrite fragment memory? Is
it doing something illegal with the print manager that might cause other
applications to crash? Could it be the Localtalk cabling? An act of god?
Thanks in advance,
Joe Murphy
Computing Analysis Corp.
DARPA IRC
NETS: jam@a.darpa.mil, jamurphy@a.isi.edu
BIX: jamurphy
------------------------------
Date: Mon, 03 Apr 1989 16:41 PDT
From: GORR <GORR@uwacdc.acs.washington.edu>
Subject: HELP! Lightspeed C
(I hope I sent this to the right place)
I receintly purchased Think's Lightspeed C for my MAC SE and I am having
problems getting the first example in the book to work. It gives
me a linking error problem. I have a MAC SE with 2 800k drives and 1 meg.
of memory...not the best stats, but according to the manual, the first
example should work...I think the problem is that the disks are set up
incorrectly (the disks are the backup copy I am useing) and in the
manual it explains how to set up Lightspeed C on a 2 drive system, but
the way the manual explains it, doesn't work...I've tried all sorts of
things, but nothing seems to work.
If you have any ideas, or I left something out that may lead you to
discovering the solution, please send Email to:
GORR@UWACDC.acs.washington.edu
Thanx...
The the first example in the manual is:
#include <stdio.h.
Main ()
{
printf("hello world\n");
}
------------------------------
Date: 1 Apr 89 05:18:41 GMT
From: Alastair Milne <milne@ics.uci.edu>
Subject: Info-Mac Digest V7 #55
>Date: Mon, 20 Mar 89 16:07:14 CST
>From: brianc@saintjoe.edu (Brian Capouch)
>Subject: Sit Story
>
...
>
> We have no trouble accessing the info-mac archives whatsoever.
>The "product mix" we're using is as follows:
>
> 1. We begin an "ftp" session and connect to the server
> using our Sun 3/150.
>
> 2. We download your .hqx files in ASCII mode; the WSMB files,
> which are predominantly .sit files, we do in tenex mode.
>
> 3. The trip to the Macs from our
>Suns is done over our Ethernet, using the Mac/IP product which we obtained
> from Stanford.
I don't know anything about this phase. I use MacKermit0.9(40) over
a phone modem from home. But a byte stream should be a byte stream,
right? Which is what Kermit in binary mode sends (unless there is some
unfortunate extra "intelligence" in a particular implementation).
>Somehow, the .sit files from the WSMR server are not recognized as sit files
>by the Unstuffer. We have tried to transfer them from the Suns to the Macs
>using each of the three possible modes built into Mac/IP (binary, MacBinary,
>and ASCII). Nothing seems to matter; the files are un-unstuffable, and
>as it turns out we aren't therefore able to download software from that
>server.
I was in just this situation. I used Art Schumer's MacSnoop to look at
the files I'd got at byte level. I found a very interesting thing: the
SIT file I wanted appeared to be embedded in an encapsulating format which
included at least the true SIT file's name, creator, and type. There was
also other binary information in the area preceeding the start of the true
file, but I wouldn't care to think about trying to decipher it.
So I made myself a small application which simply extracts the embedded
file, presenting the recovered name, creator, and type in a dialogue box,
in case you want them changed. This does create a StuffIt archive, and
Stuffit does recognise it. It protests it -- presumably there is more to
the file than just the archive, and the undecipherable information
probably tells, among other things, just where the archive ends. But
despite the protest, StuffIt sees and uses an intact archive. This has
let me obtain a number of SIT files from simtel20 -- which is very
fortunate, because, as I'm sure you know, they have a massive archive.
I've sent a note to Robert Thum (simtel's Mac archivist) about this, and
he has been very obliging; but he doesn't recognise the problem, or at
least, hasn't done so far. My description astonished him. To paraphrase
his response, ftp should download exactly what's in his archive: no more
and no less.
>... since (according to them) the "tenex" mode is the proper ftp mode
>for transferring these files.
It certainly was for me. Before I remembered to use "tenex" in ftp, I got
*REAL* garbage, interpretable as absolutely nothing.
If people are finding this necessary, I will be happy to send in the
little application for people to use as they want. It's crude, but it
gets the job done. Considering the great good I've had out of the
various archives so far, I'd be pleased to make a contribution in return.
My question about the whole business is this: how is a program like Kermit
supposed to know the creator and type with which to create a file it's
receiving? (I suppose the same question may apply to Mac/IP.) Unless
both ends are prepared to use attribute packets, what is to tell MacKermit
the type and creator? And since I first ftp the files from Simtel into a
UNIX directory (specifically DYNIX running on a Sequent Balance), and
UNIX doesn't keep file attributes, how could UNIX Kermit find those
attributes to send (supposing that it supports attribute packets)?
(It seems almost impossible that the encapsulating fields could in fact by
an attempt at attribute packets: apart from the fact that they're supposed
to precede file-header packets in the Kermit protocol, and are therefore
unlikely to be available for even a buggy Kermit to embed in the file,
they should be almost entirely printable characters, and these are almost
entirely binary.)
Alastair Milne
------------------------------
Date: Thu, 30 Mar 89 14:30:05 PST
From: lauac%QAL.Berkeley.EDU@jade.berkeley.edu
Subject: Leprechaun Demo -- part 1 of 10
This is the latest version of the Leprechaun Demo as of March 30 -- 2.5.6.
It is in a Stuffit archive.
--- Alex
UUCP: {att,backbones}!ucbvax!qal.berkeley.edu!lauac
INTERNET: lauac%qal.berkeley.edu@ucbvax.berkeley.edu
[Archived as /info-mac/leprechaun-part1.hqx; 156K
/info-mac/leprechaun-part2.hqx; 156K
/info-mac/leprechaun-part3.hqx; 142K]
------------------------------
Date: Tue, 4 Apr 89 13:01 CDT
From: John DeSoi <john@murph.tamu.edu>
Subject: Reversi/Othello Help
I'm implementing the game Reversi, or Othello if you like that name
better (I know, boring, but its for an AI project). Everything is
working ok but it seems rather slow and stupid. Does anyone out there
know of any references on computer implementations of this game in
particular? I could not find any. I really need to find an improved
static evaluation function. I saw a couple of nice implementations on
the Mac a while back; any comments/suggestions/ect. from anyone with
experience on this would be greatly appreciated. Thanks.
John F. DeSoi
Laboratory for Software Research
Texas A&M University
College Station, Texas 77843-3112
(409) 845-4306
BITNET: desoi@tamlsr
INTERNET: desoi@lsr.tamu.edu
------------------------------
Date: Mon, 3 Apr 89 23:29 EDT
From: LXC0300%ritvax.BITNET@forsythe.stanford.edu
Subject: those darn CLUT's
PROBLEM WITH MODIFYING THE COLOR LOOKUP TABLE:
The function change_clut() shown in the source listing section was
used to modify the CLUT for 256 grey levels under Finder 6.0 and it
worked just fine if each entry is protected right after modification.
The reason for doing this is because there are eight entries that are
some how "reserved" and the only way to modify their contents is by
protecting them right after modification. The first for loop generates
the 256 grey level table and stores them in the ColorSpec array mytable.
The function SetEntries() then sets mytable to be the new CLUT as
described in Inside Macintosh V. The second for loop then calls
ProtectEntry() to protect the entries from being changed.
However, under Finder 6.0.2, it is no longer possible to protect the
entries right after modification of the CLUT. The function SetEntries()
still works, but calling the function ProtectEntry() afterward no longer
protects the entries.
QUESTIONS:
1) What else needs to be done before or after modifying the CLUT?
2) Are there other ways to modify the CLUT beside using the function
SetEntries()?
SOURCE LISTING:
change_clut()
{
int i, start, count;
ColorSpec mytable[256];
unsigned short greyvalue;
Boolean protect = TRUE;
start = 0;
count =255;
greyvalue = 0xffff;
for(i=0; i<256; i++)
{
mytable[i].value = greyvalue;
mytable[i].rgb.red = greyvalue;
mytable[i].rgb.green = greyvalue;
mytable[i].rgb.blue = greyvalue;
greyvalue -= 0x0100;
ProtectEntry(i, FALSE);
}
SetEntries(start, count, mytable);
for(i=0; i<256; i++)
ProtectEntry(i, TRUE);
}
I would appreciate any comments and suggestions regarding this problem.
Lenny Chen BITNET: LXC0300@RITVAXC
------------------------------
Date: Mon, 3 Apr 89 11:10 PDT
From: Arnold_Tang@spd.3+.3com.com
Subject: TK/Solver?
I have it on good authority that when Lotus acquired Software Arts
(the guys who wrote VisiCalc and TK/Solver), they sold the rights to
some company in the Midwest. Lotus may be able to provide their name
and support information.
------------------------------
Date: Fri, 31 Mar 89 13:54:34 PST
From: DORY%ORN.MFENET@nmfecc.llnl.gov
Subject: TK/Solver for the Mac
I believe that support for TK/S for the Mac was dropped a couple of
years ago (or the owner at the time went out of business) but have seen
reference to a new version soon to appear. Reference must have been in
MacWeek or InfoWorld if not on Info-Mac, and was in the last week or
so.
Must confess that this bit of not much info at all is just an
excuse to see if I can really access Info-Mac via this route.
B. Dory Oak Ridge National Laboraatory
DORY%ORN.MFENET@NMFECC.LLNL.GOV
.
------------------------------
Date: Tue, 4 Apr 89 10:28:14 EDT
From: Tom Coradeschi <tcora@PICA.ARMY.MIL>
Subject: TOPS repeater use with KINETICS FP
>Has anyone been able to connect a localTalk network, extended through the
>use of TOPS repeaters, to a FastPath 4 successfully? We're having some
>problems. Any thoughts are much appreciated.
>
>John Jamison jamison@campus.swarthmore.edu
>Sys/Net Mgr jamison@swarthmr.bitnet
>Swarthmore College
>Swarthmore, PA 19081
>215.328.8508
>
John,
I'd recommend that you send mail to <info-appletalk@andrew.cmu.edu> on
this subject. If you wish to be put on the list, mail to <info-appletalk-
request> at the same host. This list deals with the sort of problems you
(and others) are experiencing.
tom c
Bill the Cat sez: "Remember. If some weirdo in a blue suit
offers you some MS-DOS. JUST SAY NO!"
ARPA: tcora@pica.army.mil UUCP:...!{uunet,rutgers}!pica.army.mil!tcora
-or- tcora@ardec.arpa BITNET: Tcora@DACTH01.BITNET
------------------------------
End of Info-Mac Digest
******************************