Resurrect V/AT?

[Bill Kennedy]

In article <99264 at sun.Eng.Sun.COM> plocher%sally at Sun.COM (John Plocher) writes:
>From article <196 at carpet.WLK.COM>, by bill at carpet.WLK.COM (Bill Kennedy):
[ John's agreeing he doesn't have time to be the champion, then spelling out
  what release level AT&T/Intel has for the '286 ]

>> AT&T's Simul-Task was the most stable and seamless DOS under UNIX ...
>> I'd be delighted if I could run PC 6300 PLUS UNIX on my AT clone.
>
>But the 6300+ has SPECIAL hardware to support DosMerge - The AT doesn't
>have it.

I don't think it was specifically for Simul-Task, but I'll concede that they
did remap memory using hardware and that made the job a lot easier.  The
PLUS never attempted or pretended to be AT compatible, the hardware design
preceded the AT, but that's not what I'm following up.  I appreciate John's
comments and I would have emailed, but I think that my questions are pertinent
and of wider interest than an email discussion.

>ASSUME a total of 10,000 owners of V/286.  ASSUME 30% are interested.
>
>	Year 1
>		3,000 people * $50 = $150,000
>
>		$150,000	Working funds
>			minus
>		$ 80,000	AT&T Source License for ONE CPU
>		$ 20,000	AT&T Source License for a SECOND CPU

Those certainly are going rate figures, but would they necessarily apply to
a non-product, i.e. one that they are neither interested in nor promoting?
I realize that their interest level could certainly change if they thought
that someone else was interested.  Maybe they could even become interested
in a "partial source" license, i.e. leave the kernel alone, binary license
that for distribution, and do utilities like they do the other options.  I'm
asking because I don't know.  Strictly from my point of view, I feel a lot
less territorial about my proprietary 8080 code than I did ten years ago.

>		$ 20,000	Computers for the above license
>				(4-8Mb mem, 350Mb, VGA...)

That seems high and I think that not only careful shopping, but also some
kind of realistic contribution would bring it down.

>		$  5,000	Legal Fees for incorporation and
>				Insurance
>		$ 10,000	Production and shipping costs (low)

$30 a whack for diskettes?  Maybe so if it was something other than just
diskettes.

>		$450,000	AT&T Royalties at $150/copy 
>				Remember, The current license is
>				only valid between Microport and
>				You.  Not Easter Computer and You.

OK, using the reverse arithmetic.  I suspect that $150/copy is right for
Vr3.2, but for Vr2.5?  If that was so and $30 for production and shipping,
then how was Microport surviving at $199 for (retail) V/AT with dealer and
OEM terms?  A $19/copy margin isn't very healthy.

>		--------
>	       ($435,000)	Salary   :-(

Could very well be, I was proposing some very loose figures for a reality
check.  John has been a lot closer to this than the rest of us put together.
I'll accept the numbers without discussion if we were talking about a '386
Vr3 effort, but '286 Vr2?  What were the real numbers for V/AT?  I'll guess
or speculate that Microport's money woes started with the '386 effort and
that they were moderately healthy in the '286 world.  I'm also aware that
AT&T changed a lot of the rules when they found out that they could make
money selling UNIX, but did they change across the board?  All versions and
platforms?  If so, I'll shut up, I don't know, that's why I'm asking.

>> owners who aren't on the net (could we buy Microport's mailing list?).
>
>	Call Jim Brain at (408) 441-0140 and ask him.

I will, thanks for the contact info.  Maybe the first step is to see what
needs to be done to find out if anything can be done.

>I hate to be a pail of cold water, but you really need to figure out the 
>costs, determine the number of customers, and THEN set a price.  Not the
>other way around.

Perhaps in textbook market surveys and writing business plans, but if you
have insufficient resources to acquire any kind of market share at all,
then you don't even set out on the undertaking.  If, for example, fewer than
100 people respond to the original query, I'll suggest that there is not
sufficient support to proceed.  That would indicate to me that by my first
metric (1 response = 3 potentials) we've fewer than 10% of John's gloomy
(but maybe realistic) example.

Not trying to end run Jim Brain, but does anyone have any idea how many
V/AT licenseees have access to usenet?  Any notions from BBS usage or
anything like that?  I'm curious because it would give us some idea of
whether or not sufficient souls exist to pursue this.  I know that some
like Rick Richardson have just given up, I sympathize, I did the same thing
on V/386.  We still might be able to harvest Rick (nothing personal here,
his article was in the same batch with John's) and other like him for some
other purpose, e.g. they have something they'd like to sell to a '286
platform.  In round numbers I wonder how many are out there.

Thanks to John for giving us the first authoritative figures to look at.
Thus far, all we've had is wondering and speculation (mostly over my
.signature).
-- 
Bill Kennedy  Internet: bill at ssbn.WLK.COM
                Usenet: {texbell,att,killer,sun!daver,cs.utexas.edu}!ssbn!bill