Path: utzoo!mnetor!uunet!amdahl!ames!ncar!gatech!mcnc!decvax!mandrill!neoucom!wtm
From: wtm@neoucom.UUCP (Bill Mayhew)
Newsgroups: comp.sys.att
Subject: Re: Reports on Trailblazer modems (especially with 3B2 and IBM PCs)
Message-ID: <1084@neoucom.UUCP>
Date: 8 Apr 88 14:34:14 GMT
References: <179@beattres.UUCP+ <1072@neoucom.UUCP+ <1451@bigtex.uucp>
Organization: Northeastern Ohio Universities College of Medicine
Lines: 55
Keywords: Trailblazer, info, line noise
Summary: Infoworld ratings

In article <1451@bigtex.uucp>, james@bigtex.uucp (James Van Artsdalen) writes:
+ 
+ Magazine reviews in general are of dubious value.  InfoWorld did a review
+ of several 80386 machines last fall, including the 16MHz *static* RAM,
+ no wait state, no refresh, PCs Ltd 386.  They rated it in CPU performance
+ slower than three or four dynamic RAM designs.  I never figured that out.
+ To be fair, there was no evidence they were favoring heavy advertisers there.

Infoworld seems to suffer from the "left hand doesn't know what the
right hand is doing syndrome".  Take for instance the columnist
Steve Gibson.  Gibson has made quite a few outrageous comments
about computer architecture, etc.  Lately, he recommended
formatting a non RLL-certified disk drive with an RLL controller,
AND to make matters worse, recommended formatting reserved tracks
on the drive.  Apparently, the columnists are separate from the
review panel, and the columnists can say anything, no matter how
silly.

What upset me was that Infoworld gave the AT&T 6386WGS a rather bad
review in the '386 machine round-up.  Of course, they were just
rating the msdos performance, which would miss many of the true
features of the 6386WGS that are Unix-related.  In our test suite
composed of Bourne shell procedures, we found that the 6386WGS
perfomred at up to twice the speed of an IBM model 80 (16 MHz)
running SCO Xenix 386 2.2.  Naturally, the difference is very
dependent on the flavor of *nix running.  Most if not all of AT&T's
Unix is native '386 code, while most, of Xenix is still 8086 code.
In particular Xenix's sed is still 8086.  Xenix's shells, vi, and
ed are '386 binaries.

I do agree with Infoworld's comments on Apple A/ux.  Dan Crabb has
some comments on pg. 43 of the April 4, '88 issue.  You do get a
lot of goodies with A/ux, but you do pay for the goodies you get.
The TCP/IP is particularly painless.  A/ux does still have the feel
like it was rushed out the door despite the fact that it was about
6 months late coming to market.  I don't really think that A/ux
would be suitable for non-gurus.  For example, there isn't yet a
shell procedure to automatically set up user accounts, everything
must be done manually, including setting file/directory ownership
and permissions -- this could be a pretty hair-raising experience
to a user that has never witnessed anything other than the Mac
OS.  There are very nice tools for translating Aux library calls
over to Mac quickdraw calls, so it shouldn't be too long before
some reasonable A/ux applications begin to seep out.  I wish there
was some way to get rid of the smiling computer face that you see
at the initial boot -- I'm not sure why, but it definitely
irritates me.

AT&T's hotline support, complained-about as it may be, is still
better than what one gets from many other vendors.

I have not seen the PC Designs '386 machine.  I have worked on
their 286 machine, which is a fairly good product.

--Bill