Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!sun!decwrl!labrea!aurora!eos!ames!hao!husc6!tut.cis.ohio-state.edu!ut-sally!utah-cs!utah-gr!uplherc!sp7040!obie!wsccs!terry
From: terry@wsccs.UUCP (terry)
Newsgroups: comp.arch
Subject: RISC is a nasty no-no!
Message-ID: <179@wsccs.UUCP>
Date: 23 Feb 88 04:08:44 GMT
Lines: 58
Summary: Why RISC is bad



	We have been porting software at the company I work for for a number
of years... consider: our primary product runs on over 130 unix machines,
all of the 286/386 UNIX cloning OS's (UNIX/Xenix/Venix/etc), Berk, SysV,
SysIII, VMS, CTOS, BTOS, MS-DOS, CP/M, MP/M, Turbodos, Whatever Apple calls
their normal MAC OS, and some nasty things you never heard of.  We also run
on one purportedly RISC system, the RT (IBM).

	Further, it is communications software which EMULATES 9 terminals
EXACTLY (no, it is NOT the simplistic 'cu', and NO, we do not do 132 columns
on a VT50) as possible given the limitations of the hardware.  This means it
has to talk to some pretty bizzarre serial/ethernet devices.  It does this
correctly.  At 9600+ baud.

	[Lest this be consired an ad, let me point out that 1) I have not
	 mentioned the product name and 2) the reader is supposed to glean
	 an idea of portablity.  It is the same code on all machines except
	 CP/M and MP/M, and they don't count]

	We have been able to port to all machines we attempted, except one
(no, not VMS... although that's a horse of a different wheelbase).  That one
was SUN's new (not so new any more) RISC machine, errorneously labeled via
an increment of a prior product, thus fooling the user into believing his
code will run.

	During our many-houred visit in this strange dimension (a usual port
takes no longer than 2 hours usually... the Harris HCX9 took 15 minutes,
including writing 5 tapes), we attempted to port 5 products, not one of
which runs on less than 60 machines.  The only one that didn't grunk (read
Segmentation fault, core dumped) was the one I have described above in great
gory detail, and that took tweaking.  We currently do not distribute this
due to an insane fear of technical support calls, although, again, it does
run.

THE REASON:  Type-casting.  You can't.  Small programs seem to, but it doesn't
work.  Bytes tend to be word aligned.  Other messy stuff.  It was not a
pretty sight (site?).  I am sure there are other problems, but geez, this is
demonstrably portable code.

	I am all for RISC machines when reasonably implimented.  My idea of
RISC is an instruction set that is sufficently small to allow the manufacturer
to call it RISC and not get sued, but sufficiently varied to allow me to go
off and have the assembler impliment enough macro's that my compiler thinks
it's running on a 680x0.  ----Oh rats!  If I can't have that, at LEAST my
portable C compiler should be.  Sun must have some good people to be able
to have ported a semblance of UNIX to this thing.

	I shudder when I hear people say "Won't it be neat when we can buy a
RISC workstation based on the Sun chip!".  EEAaauuUGGhah!


| Terry Lambert           UUCP: ...!decvax!utah-cs!century!terry              |
| @ Century Software       or : ...utah-cs!uplherc!sp7040!obie!wsccs!terry    |
| SLC, Utah                                                                   |
|                   These opinions are not my companies, but if you find them |
|                   useful, send a $20.00 donation to Brisbane Australia...   |
| 'There are monkey boys in the facility.  Do not be alarmed; you are secure' |