Comparison between SCO Xenix and Microport System V

[cmi at dartvax.UUCP]

In article <1146 at sphinx.UChicago.UUCP> ant3 at sphinx.UUCP (Butch Anton) writes:
>Hello Net-land!  OK, now that I got that out of the way, here's my
>question.  I'm in the market to get a UNIX for my AT compatible.  The two
>that I've looked at previously are SCO XENIX (which I actually worked on),
>and Microport System V, which I only got to play with a little at COMDEX.  My
>question is, does anyone have preferences, and if so, why?  I plan to do C
>development work, as well as just plain having fun with UNIX on my AT.
>

We are a software house specializing in Xenix/Unix/DOS communications
software, and we just brought up a Sperry IT (a good AT clone) under
both SCO Xenix V and Microport System V.  I can tell you a few things
we've learned.  Other more general information can be found in the
February issue of Unix/World magazine, which has a feature articale
comparing the operating systems, including IBM's Xenix also.

As far as a development enviroment goes, I would say SCO Xenix is
the choice.  The main reason is that they are using the Microsoft
cmerge compiler, which is one of the fastest, tightest compilers
I have ever seen on a small Unix box.  As an example, our software
package, which has 40 modules or so, compiled and linked in 22 min.
On a Vax or other large mini, it ususally takes at least 15 or 20
minutes, even on an unloaded system.  Under DOS on the same machine,
it takes over half an hour with the Manx Aztec compiler.  The
Microport compiler, which does have the advantage of producing
more portable COFF format code (used by the AT&T 6300+ and others
in the future), takes well over 45 minutes to compile and link
the same code on the same machine!

The difference in the code size is amazing.  The main binary for
our software is about 90K under Xenix, and about 170K under
Microport Sys V!  I don't know if they're linking in excess
library routines, or what, but that seems like a bit much.

The last point that I think is really significantly in favor
of SCO Xenix is that the Microport compiler only supports two
models, small and large, where the SCO compiler supports at
least 3 or 4 models, including a the medium model, which from
my experience is the best model to use in many cases when
developing applications for the AT.

In favor of the Microport Unix is the fact that, yes, it is
in fact a full-blown System V port.  If your code compiles
on a 3B2 or Tower, it's almost guaranteed to go together without
any errors or missing library routines (assuming your careful
in coding for portability).

So those are some impressions I've gotten so far.  I might
post some more remarks after using the two systems some more.

Theo Pozzy, Corporate Microsystems, Inc.

...!decvax!dartvax!cmi       (UUCP)
cmi at dartmouth                (CSNET)
cmi at dartmouth.edu            (ARPA)
Box A-58, Hanover, NH, 03755 (USPS)