INFO-MAC-REQUEST@SUMEX-AIM.STANFORD.EDU.UUCP (12/17/86)
INFO-MAC Digest Tuesday, 16 Dec 1986 Volume 5 : Issue 27
Today's Topics:
more on bugs in Apple software
Wedged 800k Disks
Are Audio Digitizers any good?
Re: are the audio digitizers any good?
More on Novy 68020/881 upgrade
RSG 3.0
Ready Set Go 3.0 (first impressions)
ExperProlog II
Re: ExperCommonLisp
Looking for Free (very cheap) MW/Paint/Multiplan
Generating Postscript from MacDraw
Mac+ Connectors
----------------------------------------------------------------------
Date: Tue, 16 Dec 86 09:59 ???
From: YOUNG%RCSDY%gmr.com@RELAY.CS.NET
Subject: more on bugs in Apple software
I regret the delay in sending this message which was caused by a bug --
sending to wrong address.
REGARDING SOFTWARE BUGS --
The seriousness of the problem of criticial bugs in mass distribution software
did not seem to be sufficiently appreciated by several of the respondents
in the earlier series of messages on "bugs in Apple software."
To the thousands of ordinary, non-computer oriented people who buy MACs,
a single critical bug in the software can seriously impair their positive
experience with the machine, and lead to untold and damaging ripple effects
as they express their unfortunate experience to friends.
Critical bugs, if discovered early in the life of a product, can also be used
by a competitor to destroy public confidence in a product. A developer
can rush ahead to the next "insanely great" product, but without
adequate testing and debugging BEFORE RELEASE (by professional testers
other than the initial programmers) any new product will be seriously and
needlessly jeaporized before it even gets off the ground.
If you instead attempt to rely mainly on the public to test your
product after an initial release, then it is even more important that you
have a strong and firm financial commitment to adequate software maintenance.
In its efforts to cut costs by reducing or eliminating both
initial testing, and also maintenance, Apple has made a serious strategic
mistake, in my opinion. The public will purchase a tried and true product
that is thoroughly debugged and works, over a new release of the latest and
greatest that doesn't work, every time (wouldn't you?).
For a company making large-scale software purchases in the multi-million
dollar range, a single critical bug will lead to no further purchases of
that product, and possibly no further purchases of any other product from
that developer. A company will simply switch to another product, or another
machine, to get the job done.
In my judgment, Apple has made a serious mistake by deciding to place
little or no investment in bug-finding and fixing for its mainline
software (or hardware) products. Although the competition often is no
better (and often worse) at reducing bugs, that is no excuse --
for major Fortune 500 companies to switch to Apple from IBM or DEC, it is
necessary that Apple have superior quality, not equal, in my experience.
For the general public, fixing critical bugs is in some ways even more important
than for the business market, since single users might not be aware that the
problem is in the machine and not in what they are doing, and also might not be
aware of work-arounds that others may have discovered.
- R. Young
------------------------------
Date: Tue, 16 Dec 86 11:40:55 PST
From: gunther.pa@Xerox.COM
Subject: Wedged 800k Disks
re: Fri, 12 Dec 86 16:34:41 PST
I had my 512k upgraded to a 512e little less than a
month ago, and last Sunday it jammed on a disk. It attempts to eject
it, but the two pinch rollers (grab on bottom of front of disk) don't
release completely, and therefore disk will not come completely out.
************
I had an identical problem with an equivalent 800k external drive. If
your drive is under warranty then you are doing the appropriate thing by
getting a replacement. If you are uncertain about whether or not the
dealer really is replacing the drive you might try marking it
unobtrusively in someway that you can identify. I did this when I had
my 512 upgraded to a Mac+. The tech. was shocked when I challenged him
that the machine he returned was not my machine. I had marked one of
the rubber feet on the 512 with my initials. What I had not realized
was that the front escutcheon (to which the feet are attached) is also
replaced in the upgrade.
I did manage to fix the 800k drive myself by increasing the tension on
the tiny springs that retract the two pinch rollers but I am not
endorsing the idea wholesale and take no responsibility for any hardware
hacking you may be inspired to do on your own.
Good luck
Neil.
------------------------------
Date: Tue, 16 Dec 86 11:40:47 est
From: ms1g#@andrew.cmu.edu (Mark Steven Sherman)
Subject: Are Audio Digitizers any good?
I have been considering getting an audio digitizer for my Macintosh and am
looking at the Berkeley build-it-yourself and at MacNifty's digitizer. What
concerns me about the former, is that I can't find anyone who uses the
Berkeley one who recommends it. What worries me about the latter is that the
example sounds coming in from the net are all huge in size (>100K) but short
on time (~4 measures of Nutcracker). If all the digitizers are only good for
sounds less than a couple of seconds, then I don't think they're of much
value. (I know, the problem could be the way the Mac generates sound. I have
several speculations on how to make things better, but at the moment, I'm
less interested in finger pointing than in determing what kinds of
functionality I can get.) Does anyone have sage advice?
------------------------------
Date: Tue 16 Dec 86 11:11:53-PST
From: David Gelphman... <INFO-MAC-REQUEST@SUMEX-AIM.STANFORD.EDU>
Subject: Re: are the audio digitizers any good?
Personally I think the MacNifty system is one of the most impressive
pieces of Mac add on hardware/software available. I wouldn't mess
around with the Berkeley setup since it is $50 and MacNifty is only
$100 with nice software to boot.
The reason the sound files are so large is that you are digitizing
the sound instead of generating it from waveforms. This has the tremendous
advantage of quality (have you ever heard a Compact Disk?) but the
disadvantage of producing large quantities of information. A CD can
hold 550 MBytes of information and produce ~75+ minutes of music.
Until you have lots of memory on the Mac you will be limited to a minute
or less of music on the MacNifty...the arithmetic works like this:
The sampling rate is 22kHz for the best sound quality...this gives a
frequency response up to approximately 11 kHz. The sampling size is
8 bits (instead of 16 bits for CDs)...this limits the dynamic range.
If we multiply we find 22000 samples/second x 1 byte/sample we get
22000 bytes/second. If you have a 1M machine you have ~800K free after
running the program which captures the sound so we find
800000 bytes / 22000 bytes/second = ~ 36 seconds
You can get more sound recorded at the expense of quality...the
Macnifty software allows you to use sampling rates less than the 22 kHz
so you can get more sound with a lower frequency response. You don't need
11kHz frequency response for human voice.
OK, so you can get say 1 minute of sound in 800kbytes if you want but
this means the sound file will be 800 kbtyes before any compression to
the file. Expect that the resulting file will be between 400k and 800k in
size after file compression.
The result of all this is what you might have guessed before the
estimates if you just thought about 75 minutes / 550 MBytes on CDs. Digitized
sound files are huge. One thing which is fantastic about the MacNifty
system is that you can export sounds to Studio Session which is a music
creation program like Music Construcion set, Music Works, and the one
from Great Wave Software (can't remember the name offhand) except it
can use the digitized sound files you create with the digitizer. The
result is that you can have very good sounding music you create on the
Mac. It is much better than the other programs.
David Gelphman BITNET address: DAVEG@SLACVM
Bin #88 SLAC ARPANET address: DAVEG@SLACVM.BITNET
Stanford, Calif. 94305 UUCP address: ...psuvax1!daveg%slacvm.bitnet
415-854-3300 x2538
usual disclaimer #432 applies: my employer apologies for the fact
that I have access to this net.
------------------------------
Date: 16 Dec 86 14:22:00 EST
From: <bouldin@ceee-sed.ARPA>
Subject: More on Novy 68020/881 upgrade
Reply-to: <bouldin@ceee-sed.ARPA>
What follows is the ad that Novy systems is sending out to advertise
their 020/881 upgrade for the Mac. Since there is obvious interest
on this SIG, I sent the entire text of the ad.
For the price, this looks to be a good deal. I appended some of my own
comments at the end, clearly separated from the ad text.
**********************************************************************
The 68020 Macintosh
with 68881 math coprocessor
The MAC20 is a low cost upgrade to the motorola 68020 32 bit
Microprocessor for the Macintosh 512E and Mac Plus computers. The MAC20
multilayer add-on board is available NOW from Novy Systems Inc. and may
be ordered with the 68881 floating-point coprocessor. 6881 access from
Microsoft Fortran and programs supported by the Standard Apple numeric
Environment (SANE) is included.
Most popular Mac programs have been tested on the MAC20 and show typical
improvements of 25% to 50% using only the 68020 while some applications
show up to a 100% improvement. Numerically intensive programs using the
MAC20 with the 68881 installed can expect improvemennts of up to 400%.
The MAC20 card also includes:
1.A desk accessory to control the features of the MAC20.
2. A utility to provide automatic access to the 68881 from
Microsoft Fortran (2.1 and 2.2).
3. MDS assembly languag 68881 interface example.
4. A utility to provide access from TML Pascal, Macspin, and
other packages using SANE.
Compatibility: Our testing has revealed that only programs using four
voice or freeform sound output, such as music programs, will not operate
properly with the MAC20 board. These appear to be software
incompatibilities and should be corrected in future versions of these
programs. Programs that are self-modifying may not work properly while
the 68020 instruction cache is enabled. A desk accessory is provided
that allows the cache to be easily diabled for these programs.
A partial listing of programs found to be improved with the MAC20 are:
MacPaint MacDraw Excel
Word MacChallenger Fortran (2.1 or 2.2)
Fligh Sim. MS Basic MacTerminal
MS Works ReadySetGo MS File
Pagemakeer Dollars & Sense Sargon III
Benchmarks
MS Fortran 2.2 Absoft MacFortran 020
Mac MAC20(with 881) MAC20 (with 881)
KiloWhetstones
Single Prec. 41.6 152.0 193.5
Double Prec. 18.5 108.7 178.0
Sieve 6.5 3.5 2.9
Absoft Fortran/020 is an optimizing compiler for the 68020 and 68881.
Other currently available optimizing compilers on the market which
support the 020/881 are Consulair C and the soon to be released TML
Pascal.
Introductory Prices:
Part# Description Price
MC20 MAC20 (without 881) $595
MC2081 MAC20 (with 881) $749
Installation of MC20
(desolder original 68000 and $49
insert socket)
ADD"-C" Clip on Version add $99
ADD"-R" RADIUS compatible version add $99
6888-12 MC68881rc12 chip $195
MCXX-2P Board and PROM set (w/o 020/881) $395
Clip on for Mac+ available soon
6 months parts and labor warranty.
Prices subject to change without notice.
Novy Systems Inc.
69 Ravenwood Ct.
Ormond Beach, FL. 32074
(904) 427-2358
***********************************************************************
Comments:
Looks like good price/performance since this allows about 40% of prodigy
4 whetstone performance for about 18% of the price.
Also, this is considerably faster than any MS-DOS Fortran on even the
hottest AT clones around. Computer Languages, Jan. 86 issues, benchmarks
Lahey MSDOS Fortran at 98K single precision whetstones and 83K double
precision, while Ryan-McFarland did 106K, and 89K, respectively.
Hardware was an 8Mhz Compac 286 DeskPro. On the newer 12Mhz 286 machines
this would go to about 150K single precision whetstones.
The newest wave of 16Mhz 386 machines should only (roughly) scale by the
clockrate if they are running compilers which do not explicitly generate
code for the 386 (this is a best case, since there is, as yet, NO 387
chip. The DeskPro 386, for example, actually uses a 12Mhz 80287 for
floating point). This means that the Mac with the Novy 020/881 board
and Absoft 020 code generator should be essentially the same speed as
the new 386 machines.
Also, don't throw out your MS Fortran 2.2 (or 2.1) release discs!! They
are worth $100 in an exchange/upgrade for the Absoft 68020 code
generator. That brings your net cost to $400 for the 020 code generator.
Non floating point stuff that is speeded up the most by the 68020 is stuff in
tight loops. Things with tight loops will perhaps double in speed. This should
include a lot of quickdraw, like the bit copies. It is possible that a true
9600 baud terminal emulation could result from using the Novy board.
****************************************************************************
I do not have any commercial interest in Novy systems. I just think
there is a widespread interest in these upgrades, especially if Apple
makes the colossal blunder of having no 020/881 upgrade for the Mac+ along
with the Jan. announcements.
------------------------------
Subject: RSG 3.0
Date: Fri, 12 Dec 86 09:23:03 -0500
From: meltsner@athena.MIT.EDU
I've been bashing away at Ready,Set,Go3 for the last couple of days.
[MacConnection was not ready to ship it when I ordered -- the first
time ever!] It is pretty nice, handles all of my old documents (including
scientific stuff with subscripts and superscripts, footnotes, etc. from Word).
The manual is funny -- a "magazine" on how to use RSG. There are a
few bugs with redraw [specifically, if you muck with baselines on a
very short text block, the redraw leaves gunk around -- fixed by
scrolling to some where else], but on the whole it is fast and fun.
The real-time H&J even works.
Some of the graphics are a bit messy. Try doing corners on a fancy
border.... On the other hand, it does have thin rules (~1/2 point),
and a nice cropping/resize system. The automatic text runaround of
graphics is especially slick.
The Postscript windows are too primitive. What you get is literally
just a way of sending raw PS to the laserwriter, along with the rest
of your layout. No scaling, no translation of coordinates, nothing.
This means it becomes tough to just drop in a PS file into a block,
and have it print in that block's position and size automatically.
You have to look it up in the block specifications and edit your PS to
match.
RSG reads TEXT and PICT files, MacWrite and Microsoft Word documents,
and appears to be very Switcher and DA compatible. It uses the
standard Apple Laserwriter driver and Prep file.
Ken (meltsner@athena.mit.edu)
------------------------------
Date: Fri, 12 Dec 86 10:05:35 PST
From: chuq@Sun.COM (Chuq Von Rospach)
Subject: Ready Set Go 3.0 (first impressions)
Ready Set Go 3.0 has finally hit ComputerWare, and last night I had the
time to sit down and play with it a little bit. In general, I'm VERY HAPPY
with the product -- after fighting my way through MacPublisher II for the
last 6 months, I can't tell you how much easier this program is to work
with and get things accomplished.
The program is rather easy to use, but not always intuitive -- read the
manual first, and keep it handy for the first few sessions.
[editorial side note. The design of the manual is a great testimonial
for the program. It's a very complex, clean document, done as
a 65 page magazine on glossy paper. On the downside, they tried
a little too hard to show all the neat things RSG could do, and
I find the document a little too busy on the design side. I also
worry about the long-term prospects of protecting a glossy magazine
from the trials and tribulations of my desk -- a heavier card stock
would last much longer under normal usage]
On the program side, RSG has a lot of neat features:
o Postscript windows. Yes, all you need to do is create a block of
text, click a button, and type in a postscript program. Think of
the possibilities. (like, wow!)
o real-time hyphenation. Yes, you can toss an article on a page,
turn onhyphenation, type, and watch the entire screen redraw with
the changing hyphenation -- fast enough that it doesn't really
bother you.
o attribute changing. You can change attributes on any size block
of text -- from a character to the entire document. This includes
justification, font, style, size.
o Text wraparounds. Build a block. Drop it on the page. Watch
everything squirm to get out of the way. I think I'm in lust...
(it works, wonderfully!)
o full word processing features. search/replace and a spelling
checker.
o it doesn't brew coffee for you. yet.
Overall, I'm very happy, with a few minor reservations and one bug. It
does everything I want it to do (that I know how to do, so far) and the
results of the test designs I threw at it last night were rather wonderful.
Printing, by the way, at least to a laserwriter, is quite fast.
The few gripes:
o The word-processing isn't as powerful as they would like you
to believe. They claim you could do ALL your writing in RSG3,
and throw the rest away. No, you can't -- it would drive you up
the wall. But, it is more than useful if you're putting together
short filler pieces or hacking an article into shape. It's good,
but it could be better.
o There are no decimal tabs. There are up to 9 left, right, centered
or justified tabs, however.
o There is a bug in the search/replace algorithm. choose 'search'
and search for '"'. When you find it, replace is with
shift-option-[ (the open double-quote). When you go to search again,
the 'change' command has also changed the search field, making
a 'find next' operation impossible if you use the 'aXnge' option
(instead of just typing in the new text).
o the spelling checker is very weak. It's MUCH better than none
at all, but don't plan on tossing Thunder! or Spellswell yet. You
can't modify the RSG dictionary, either to add or delete. You
CAN set up a User dictionary, but it is tied to the application
(where RSG lives, it lives) and not the document, so you can't
make spelling dictionaries on a document by document basis. foo.
o The index is weak -- for example, it doesn't tell you the page
that has the information on the command to put the page numbers
on your pages. The documentation overall isn't bad, but it could
have been better organized.
In general, it is the best thing since slightly stale bread. Lots of fun,
and it makes MacPublisher II look like teco.
As I find more, I'll let you know.
chuq
16-Dec-86 05:51:42-PST,592;000000000001
------------------------------
Subject: ExperProlog II
Date: Tue, 16 Dec 86 08:54:28 EST
From: Jonathan K. Millen <jkm@mitre-bedford.ARPA>
Has anyone used ExperProlog? Is it as good as it sounds? How about
its development environment? Its toolbox interface? Is it fast?
-Jonathan Millen
jkm@mitre-bedford
...decvax!linus!security!jkm
------------------------------
Date: Tue 16 Dec 86 10:30:00-CST
From: CS.RICHMANN@R20.UTEXAS.EDU
Subject: Re: ExperCommonLisp
In the 15 Dec 86 info-mac digest Bill Roberts asks:
>Do we get ExperCommonLisp for free or do we get to pay a nominal
>upgrade fee of $750 (and that does not include the file compiler)?
Like Bill, I was confused by the ExperTelligence marketing strategy
of plan B upgrades offered to present owners of ExperLisp. My
questions, however, were answered by the arrival yesterday of an
_eight_ pound package of software and documentation from
Expertelligence. Here is what I found:
-- Customers who purchased the plan B upgrade to ExperLisp 1.5 will
receive the complete ExperCommon Lisp package, including file compiler
(for producing double-clickable, stand alone applications) _without_
additional charge. The software includes the common lisp listener,
incremental compiler, file compiler and a run time kernel for inclusion
in your own applications. A few examples are also included. The
documentation is in three indexed volumes. A "dead code" optimizer
is discussed in the documentation, but not included in the package.
The software comes on three 800K diskettes and is not copy protected.
The following are my unsolicited opinions on the package:
-- The development environment works only on a MacPlus, but the
applications you develop should work on a 512K Mac with either
the new or the old ROM's. You will appreciate a hard disk and
a memory expansion board (like the MaxRam, etc.). This is the
largest program I have running on my Macintosh -- the
kernel alone is 339K and the compiler is 363K.
-- Compilation is slow. It takes two or three minutes just to load
the compiler and compilation is painful compared to the LightSpeed
compilers for C or pascal. You can, of course, use your compiled
programs in the listener environment without any additional work, but
to produce a stand alone application, you must use ResEdit to "paste"
the compiled code into a lisp kernel.
-- The compiler supports a full implementation of common lisp
with object oriented programming extensions. The support for the
Macintosh toolbox is very complete and well documented. I compiled
a moderately large expert system written in common lisp on the
Dec 20 and only had to change one thing. The item that needed
changing was a holdover from Experlisp 1.5 in that symbols
beginning with a dollar sign are interpreted by ExperCommon Lisp
as hexadecimal numbers.
-- Keeping in mind that I have only had the package less than
24 hours, I do feel as though ExperTelligence has solved the
problems with bugs that plagued their earlier product. The
development environment has a nice debugger, and I have had
no mysterious crashes.
-- The editing environment is "Mac'ish" with a few EMACs like
extensions such as parenthesis matching, automatic indentation
and cursor movement with the keyboard. The editor is based on the
text edit toolbox, however, and is limited to files of 32K or less.
(Doesn't this seem cheap on a $995 software package, even though
in practice it does not represent a significant limitation?)
-- As to the price, I agree with Bill that it is exorbitant for
an individual to pay this much. I also suspect that there will
be a steep cost to upgrade to a 68020 machine code version next year
when the "paris" and "alladin" class machines hit the market. But,
look at what Golden Common is charging for their 286 developer
version of lisp. Apparently, defense contractors and other
industry types are paying these big bucks for lisp software. I
also don't think ExperTelligence is making millions from this
package -- they have given me _excellent_ support over the last
two years, and I am not a big bucks customer.
Disclaimer: I have no interest or involvement with ExperTelligence
other than as a satisfied customer. I will be glad to answer
questions to the best of my ability. Direct them to
cs.richmann@r20.utexas.edu or James Richmann, 4906 Caswell Ave.,
Austin TX 78751.
------------------------------
Date: Mon, 15 Dec 86 22:41 EST
From: Mark.Sherman@A.CS.CMU.EDU
Subject: Looking for Free (very cheap) MW/Paint/Multiplan
I will probably be giving a small set of lectures to high school
students on general computer science. I would like to have these
students play with a word processor, a drawing program, a spread
sheet, some assembly language stuff (a debugger would probably do)
and a Lisp (preferably object oriented). I have a very modest budget
for 100 students, so getting N copies of just about anything is out
of the question. In olden days, I could use MacWrite/MacPaint with the
knowledge that every machine used was purchased with it, so I could
just copy them and leave them around. That is no longer true. A quick
glance at the <info-mac>archives and at several Public Domain
catalogs doesn't offer much hope. If necessary, I will replace Lisp
with Smalltalk, a word processor with a text editor, use one the
trivial drawing programs (that come in TML or sommeplace else) for
the MacPaint and the built in debugger of the Mac+. I am checking
into XLisp (though the archives look old to me -- it's in HCX
format!). I saw nothing even coming close to a public/shareware
spreadsheet. Have I missed some obvious program for any of
these needs?
-Mark
------------------------------
Date: Tue, 16 Dec 86 8:00:37 MET
From: Francie Newbery <newbery%germany%germany.csnet@RELAY.CS.NET>
Subject: Generating Postscript from MacDraw
[I realize that this is not quite the appropriate forum for this
type of "request", but I was hoping that you would be willing
to oblige anyway.]
We would like to be able to take the Postscript generated by
MacDraw and include it in our non-Apple textformatting programs
(i.e. ditroff, TeX). I understand that this can be done, but
with care. I read in info-postscript that this has previously
been discussed in info-mac. I was hoping that you would be
able to mail me an appropriate article which would describe
1. How to get hold of the Postscript file in the first place
and
2. What kind of things do you have to watch out for.
i.e. Do I need to have some kind of different header file;
Something about Apple Postscript using 8 bits causing
difficulties; bitmaps causing problems, etc.
Thanks in advance,
Francie Newbery (newbery@germany.csnet)
------------------------------
Subject: Mac+ Connectors
Date: 16 Dec 86 14:07:42 EST (Tue)
From: prj@pm-prj.LCS.MIT.EDU
Does anyone know of a source (particularly in the Boston area) for the
round Mac+ connectors? NOT the complete cables. We want to make our
own and just need the actual connector hardware.
Thanks for any information.
---Paul Johnson
------------------------------
End of INFO-MAC Digest
**********************