z (04/28/83)
Funny someone should ask about UNOS. In one of their ads they proclaim that they are v7 compatible. Bull! They may wind up being v7 compatible with their next release but they are far from it now. The main problems are that their stdio is different and there are no catchable signals. Their stdio package has a different idea of file descriptors and FILE pointers. As their documentation stands presently they list all the unix library subroutines but only implement as small fraction. Most of the unix commands are there with different names. There is a clumsy alias facility to allow unix commands to be issued. They don't seem to have an "enviornment" and thus no such things as run paths. Some of the commands have a non-unix flavor as far as option select is concerned. It looks like they haven't got all the bugs out of their C compiler. At least the pre-processor fails for certain invocations of macros. They said their 3rd release will be v7 compatible. It was due in July of this year but may be postponed because of "loss of key personnel." It looks like if you are doing original real-time programming that it may be a good system. Unfortunately it does not compare favorably with a "real" unix system. //Z\\ Jim Ziobro
mjb (04/29/83)
Ever see those UNOS ads that claim Chrles River has sold more UNOS licenses
than WECo has sold UNIX licenses? Ever wonder who bought them? Well I found
out that Tandy bought 30,000 licenses for a UNOS port to the TRS-16 (I think
they also sold ~120 UNOS licenses for CRDS's own machines). Anyhow, Tandy
ended up scrapping UNOS in favor of XENIX.
Mike Braca, Brown CS, {decvax,ihnp4}!brunix!mjb, mjb.brown@udel-relay
PS: UNOS has no UUCP!guy (04/30/83)
>From what some CRDS people said, their system calls don't return the error
status in an external int called "errno", and the error numbers are different
anyway. Furthermore, they decided that nobody uses groups anyway, so there's
no GIDs anywhere.
According to somebody from Microsoft who played with one of their systems,
their shell isn't UNIX compatible either.
And these guys are advertising themselves NOW as being UNIX-compatible. Well,
I wish people who decide that they're smarter than Ken and Dennis and the
entire USG all put together would AT LEAST build their systems so that things
that used to work on UNIX work on their wonderful new all-singing-all-dancing
system (or if they're offering an "enhanced" version of UNIX, at least
"enhance" it so that stuff that used to work still works). If you end up
working with more than one type of machine, you very quickly get tired of
the sort of anarchy that often prevails with UNIX implementations. When
the /usr/group Standards Committee comes out with its standard (yes, I know
it's taken a long time; I'm a member of the committee, and it just takes a
long time to thrash out all the issues, especially with ~20 people), maybe
some of this will clear up.
There are representatives from CRDS on the committee, so there's hope for
UNOS yet. Also, finally, there are representatives from Bell Labs, and the
latest note says that Bill Joy is now on the committee; most of the major
UNIX OEM's are also represented. It seems broadly-based enough that the
standard (and future extensions) will, with any luck, provide what most
people need.
I may submit an article to net.news.group suggesting a newsgroup for
discussing the /usr/group (or whatever it's being called today) standard
and other standards, either for UNIX and UNIX-like systems or standards
in general.
Guy Harris
RLG Corporation
{seismo,mcnc,we13,brl-bmd}!rlgvax!guy