muller@munnari.oz (Paul Muller) (11/30/88)
I know this is asking for comp.unix.wizards.flame, but I am rather confused.
I have been talking to people in the industry and to collegues over the
advantages of SCO Xenix vs Microport Sys V (AT/386), which is better, the last
comment I recvd from a friend (who runs Xenix) was that Microport was CRAP!
He cited the point that the ttyx drivers are stuffed and that the whole OS is
a snail! This is rather odd as a conflicting comment suggested that M/Port
runs rings around Xenix on any PC?!?!
I would appreciate some expert (read:user) opinion on this as I have
to help manage a system that will run a file intensive retrieval system and
with around 6 users I need something that can hack the pace and save my bum.
I apologise for the contravertialnous (spel?) of the topic, but I really
like to know what I'm getting myself into (assuming it makes it through finance
Thanks in advance,
Paul Muller
+61 3 743 5930
(muller@munnari.oz.au)davidsen@steinmetz.ge.com (William E. Davidsen Jr) (12/02/88)
I can only speak to the 386 side of Xenix vs. Microport, and of course
based on my own experience, but I have used both quite a bit, as a user
and administrator.
Early versions of MP had problems with the serial drivers. We reported
it, MP blamed our software, then provided fixes. After 3rd fix the
system no longer crashed when serial input was used, but it drops data.
There are later fixes, but that user gave up.
286 versions of MP don't support as many memory models as Xenix. They
don't cross compile for MS-DOS as a standard feature (if you don't need
it, so what).
The Xenix compiler is not PCC and has its own set of bugs. I don't think
there are as many as MP, based on the programs I've run. Over all of the
systems and applications programs I've measured, the Xenix compiler
seems to produce faster code. MP will be slightly faster on some stuff,
while Xenix will be slightly faster to vastly faster on other programs.
Xenix uses a more V7ish tty control structure, and while the V.2 stuff is
there I don't know if it works. With 2.3.1 Xenix provides a /dev
structure which is a superset of the V.2 structure. Xenix provides HDB
uucp which will be in V.4. Much better security and lower administration
effort. They provide some useful tools to help with setup, etc.
The number of software packages available today for Xenix is about 5:1
greater than MP. In a year or two Xenix and V.x binaries will run in
both systems, but right now that's the case. Xenix claims to run V.2
binaries, and does at least some. None of the people I know with MP have
a version which will run Xenix binaries.
My experience with Xenix support has been mixed, but better than with
MP. Xenix seems to come out with fixes in new versions and has MANY fix
packages you can order for free, MP has a BBS. Xenix supports
Trailblazer, I have never seen a MP serial port read data correctly at
19.2.
Sorry this isn't a nice uniform coverage saying "X is better." You will
hear from people who favor each type. I would suggest that you look at
their experience with both and decide if they are owners of one and
casual users of the other, or experienced in both. There are lots of
people who are happy with each. Both are pigs if you don't give them
enough memory!
Money where my mouth is dept: I tried Xenix, MP, and INteractive, and
bought Xenix with my own money.
--
bill davidsen (wedu@ge-crd.arpa)
{uunet | philabs}!steinmetz!crdos1!davidsen
"Stupidity, like virtue, is its own reward" -memuller@munnari.oz (Paul Muller) (12/03/88)
It seems I bit off more than I could chew! I thought I'd get one or two responses but the reaction has been helpful, but I will have to call thenough enough. I will read through all the mail I have received, try and sort out the overlap, and try to prove some MORE conflicting claims either way..... I now have a file of contacts and that will help the procdedure greatly, thanks yet again, Paul. "now where did I put that manual", Paul Muller.
shocking@physiol.su.oz (Stephen Hocking) (12/08/88)
In article <2596@munnari.oz>, muller@munnari.oz (Paul Muller) writes: > > I know this is asking for comp.unix.wizards.flame, but I am rather confused. > > I have been talking to people in the industry and to collegues over the > advantages of SCO Xenix vs Microport Sys V (AT/386), which is better, the last > comment I recvd from a friend (who runs Xenix) was that Microport was CRAP! > He cited the point that the ttyx drivers are stuffed and that the whole OS is > a snail! This is rather odd as a conflicting comment suggested that M/Port > runs rings around Xenix on any PC?!?! My experience is (2 Yrs with Xenix 286, 1 1/2 yrs with Uport V/AT) is that the bugs are simply different. Xenix runs faster with small amounts of memory (seems to swap less, basically) but Uport overtakes it as the memory size gets larger. The tty drivers under Uport were stuffed, but that has been fixed OK with the later releases. I've been using it from 2.2 through 2.4 and the improvement has been immense. With the development system, the Xenix 386 compilers were pretty sad, but Uport had it right almost first up. It did frustrate me at first, but at least I dont see segment fix up over flows.... Xenix has better support for perverse hardware, although Uport has just added RLL & ESDI support. Mail me Paul, I'm in the same country... Stephen -------- "You are surrounded by piles of messy nappies, all alike" Stephen Hocking ACSnet shocking@physiol.oz UUCP ...!uunet!munnari!physiol!shocking
vandys@hpcupt1.HP.COM (Andrew Valencia(Seattle)) (12/10/88)
/ hpcupt1:comp.unix.wizards / shocking@physiol.su.oz (Stephen Hocking) / 5:22 am Dec 8, 1988 / > With the development >system, the Xenix 386 compilers were pretty sad, but Uport had it right >almost first up. That has been my experience with the '286 products, too. Xenix is higher quality, but consistently harder to port to. It just seems to come up with more reasons to break things.... Just opinions, mind you! Andy Valencia ...!hplabs!hpisoa1!vandys