dmkdmk@ecsvax.UUCP (David M. Kurtiak) (05/04/88)
The following are reviews that I have received regarding the POWER-C inquiry
I made earlier last week. Thanks to all who replied! Looks like a package
that may be worthwhile to say the least.
----
David M. Kurtiak
UNC - Greensboro
UUCP: dmkdmk@ecsvax.UUCP {decvax,rutgers,gatech}!mcnc!ecsvax!dmkdmk
Bitnet: DMKDMK@ECSVAX.BITNET (mail ONLY)
Internet: dmkdmk@ecsvax.uncecs.edu
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Date: 28 Apr 88 00:53:31 EDT (Thu)
From: rjs@moss.att.com (Robert Snyder)
Organization: AT&T Bell Laboratories, Whippany NJ
A friend of mine bought it and liked it a lot because it did a good
job with floating point calculations and he was having all sorts of
trouble with Turbo-C. So on his recommendation I bought it and tried
to compile nethack (might as well give it a good test). It didn't do
too well because it keeps its symbol table in one 64K segment, so
runs out of memory in a hurry. I called up Mix software and they
said they have a new version which uses the "far heap" for the symbol
table thus providing much more capacity. They charged me another $10
for an upgrade. I bought the compiler in January.
The compiler rejects some of the more esoteric syntactic constructions
that show up in hack, like
char line[MAXLINE], *pl = line;
It claims line is undeclared for use to initialize pl. You have to break
this into two statements. There's a few others, but all have simple
work-arounds. (Yacc output must be edited before it can pass syntax).
Anyway, I don't think it's production quality yet, but if you need better
floating point than Turbo-C can provide you might be interested.
It comes with a big manual, but the section on the C language seems to
be a standard text, not specific to the POWER-C compiler. For example,
it doesn't tell you how big int, long int, and short int are, but gives
a code example to try on your compiler to find out. The section on
the library calls seems pretty good but the rest is somewhat skimpy.
Robert Snyder
moss!rjs
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Date: Wed, 27 Apr 88 15:51:37 PDT
From: jpl-devvax.jpl.nasa.gov!jackm@rutgers.edu (Jack Morrison)
Organization: Jet Propulsion Laboratory, Pasadena, CA.
I'm waiting for reviews too. A recent mention appeared in INFO WORLD (or
was it PC WEEK) - the reviewer said it failed to work if you use
recursion, so he was waiting for a "fixed" version before running any
benchmarks. Sounds a bit scary, although source code to all those library
functions might be worth the whole ticket...
--
Jack C. Morrison Jet Propulsion Laboratory
(818)354-1431 jackm@jpl-devvax.jpl.nasa.gov
"The paycheck is part government property, but the opinions are all mine."
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Date: Wed, 27 Apr 88 18:16:32 PDT
From: harvard!seismo.css.gov!megatek!hollen@gatech.edu (Dion Hollenbeck)
This package IS THAT GOOD and it is true if their first release was
any indication. The first release had only a few shortcomings.
1) Did not create .obj files which could be linked, instead, created
.mix files which were p-code. A p-code interpreter was put at
the front of the .com file and these .mix files conacatenated.
Was possible to combine with assembly with utilities they provided.
2) Only supported small model (combined 64k code, data, stack).
The documentation is EXCELLENT and chock-full of examples (really good
for the beginning C programmer). Their technical support staff was
very helpful, but seemed not to be really technical (the perspective I
view from is that I am a Senior Software Engineer and program real-time
embedded systems in assembly and C for a living).
The debugger is one which all others should be judged by.
The original package was sold for $79.95 about two years ago, but since
release of the improved model, went as low as $19.95. I believe the
first release of POWER C (sans debugger, which is why I stayed away;
I feel a compiler is worthless without a source level debugger) was
also $19.95.
From what I hear from their technical support people about a year ago,
teth current release supports real .obj files and is compatible with
Microsoft object module format; it was also going to support all the
common memory models.
Hope that gives you an idea of whether to go for it or not. If you
do, please let me know how the product is now.
Dion
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Date: Fri, 29 Apr 88 08:18:16 PDT
From: marlin.nosc.mil!herman@rutgers.edu (John W. Herman)
Organization: Naval Ocean Systems Center, San Diego
--
I've used POWER-C a little bit, but I'm not an accomplished C programmer.
Documentation is first rate and may be the best feature of the package. I
had a problem with function calls that is probably my fault. I got a strange
glitch doing a Hilbert transform of some data and I was unable to track it
down. The debugger wasn't available until recently and I haven't used it.
Overall reaction, I think it is a good to excellent value for the money.
--
John Herman ARPA: herman@marlin.nosc.mil
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
<end of text file>