DOS vs Unix application quality (was Re: Interactive and me )

[John G. DeArmond]

bill at ssbn.WLK.COM (Bill Kennedy) writes:

>>to produce a package that meets the quality standard of the DOS world from
>>even as far back as 5 years ago (you know, things that generally work like
>>they should, documentation that bears some resemblance to what's shipped,

>John, I don't think it's entirely fair to equate the DOS world with Interactive
>UNIX or any other UNIX for that matter.  The "quality standard of the DOS world"
>is pretty sad until you have curried and combed out the pearls.  DOS runs and
>runs well on specific platforms, UNIX crosses all the borders.  I whole-
>heartedly agree with your conclusion (ditch Interactive) but I disagree with
>the comparison.

Let me clarify a bit here.  When I refer to the DOS world, I do not mean
the program loader actually called DOS, I'm refering to the applications
that get loaded.  And since each application must have major portions
of an OS embedded in it, this is a fair comparison.

Most any DOS application of substance (translate: you gotta pay 
$ubstantial bux for it.) meets a minimum standard of quality that is
vastly better than what we put up with in the Unix world.  If the
Docs are incomprehensible or poorly organized, if the program crashes
or does not perform in minute detail in accordance with the docs, 
it gets flamed in the press and people won't buy it.  

Yet we in the unix world continue to tolerate shitty documentation, 
utilities (the bundled "applications" of Unix?) that are broke,
installation procedures that don't work, and vital information
that "you just gotta know" in order to make the systems perform.
For example, I have  yet to see a Unix box whose line discipline
behaves exactly like the termio man page says it should.  Async I/O is
the place that my code invariably has the most system-specific #ifdefs.
There is no excuse for this.  Unix has been around too long.  And unlike
the DOS applications vendors, Unix vendors such as ISC did NOT have to
actually write the software.  All they've done is port and optimize (or
in some cases, break) the AT&T supplied code.  Not to  denigrate the
good work done by ISC and others, but the time normally used to create a
product should have gone into documentation and stabilization.

I look at what a few vendors such as HP, Sun or NCR have done with
documentation and it makes me want to cry.  NCR, for example, not 
only supplies a tuning guide that explains kernal parameters and how
each one affects system operation, they also supply a separate book
of error messages with explanations as what to do to remedy the problem.
Why can't we have  that kind  of quality in the PC Unix world?  I
love nothing more than to knock some huge piece of  computing iron out
on the street with a tiny little PC that runs circles around it.  The
recent work of ISC and others is making that task more difficult.

<Climb off the soapbox for a moment>


>[ more DOS comparison deleted...]
>>installation, while a bit more glitzy than 2.0.2, had all the same 
>>old problems and still required me to whiz around with vi editing this

>Here's where he scares me even farther away from 386/ix v2.2.  John says that
>all of the nagging old problems are there, or at least enough of them to
>feel like they're all there, but some new layers of distraction have been
>added.  I don't want glitzy or exciting installations, I want dead, dirt simple,
>thoughtless effort.  I don't want to have to figure out some nifty script or
>find my way into some programmer's head to figure something out; just lay it
>down on the disk and give me a login prompt.  No thanks, the excitement doesn't
>enchant me.

Just to be fair,  a couple of the glitzy features are actually worthwhile.
There is an interactive file system partitioning utility that gets 
run as part of the installation that is very nice.  It operates spreadsheet
style and allows you to work  out disk partitions while showing you the
results of  what you've  done.  This is nice.  The bad part is that it
seems (I say 'seems' because I've  not had time to search  for it too
much) to exist  only on the installation floppy.

On the other hand, they nuke themselves  (and us) with arbitrary changes
in the installpkg routine.  I spent the day and most of the night
friday working with a vendor of IBM connectivity hardware/software
trying to get it to install under 2.2's installpkg.  It  appears that
ISC has made  somewhat arbitrary changes to the scripts  that render 
2.0.2 scripts obsolete.  Once we got the scripts fixed, the 
application  went right in.  What they've done is force vendors to now
have  BOTH versions  up and running on separate machines  in order to
support both.  Wonder if it would have been possible to have kept the
old stuff, maybe named "oldinstalpkg" or something like that?  nah...

>Thanks for the follow-up John and sorry for the bulk.  I had to sit on the
>article for a few days to make sure that I was making a business decision
>and not just reacting emotionally.  Nope, I run a business, I depend on
>happy customers, I must maximize my productivity to make a profit, and
>Interactive keeps building obstacles.  I don't fabricate or imagine them,
>Interactive constructs them.  They've got the last of my bucks and join the
>list with Microport, Santa Cruz Operation, and American Telephone & Telegraph.
>Jeezus the list is shrinking as firms become prosperous!

Yeah, me too.  Why does it have  to be that way.  Why is it that the
first  thing  a newly successful company does is forget who and what
made them successful?  Why is it that if you don't represent 7 digit
annual sales  to them, you are to be ignored?

As  a side note,  I've  gotten as much mail in response to my original
post as for anything I've ever written for the net.  My mailbox 
runneth over!  And not a single  person has disagreed.  I've  tried
to  reply to all these people but sometimes  our mail does not work
too good.  I am urging everyone who  has responded and even those
who  have not to let ISC and SCO know how we feel.  Call, write, or 
just post (they are watching.) and speak your mind.  A similiar
campaign killed copy protection in the DOS world and it can do it here.


John

-- 
John De Armond, WD4OQC  | We can no more blame our loss of freedom on congress
Radiation Systems, Inc. | than we can prostitution on pimps.  Both simply
Atlanta, Ga             | provide broker services for their customers.
{emory,uunet}!rsiatl!jgd|  - Dr. W Williams |                **I am the NRA**