[TUHS] What UNIX Artifacts Are Still Missing?

[clemc@ccc.com (Clem Cole)]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 3160 bytes --]

On Wed, Dec 6, 2017 at 9:33 AM, Paul Ruizendaal <pnr at planet.nl> wrote:
>
> ​A​
> lthough Xenix predates Venix, I'm not sure it predates it on PC hardware.

​It really depends how you count.   I was there at time when AT&T was
negotiating the replacement for the V7 license with 10 of us (the 10 firms
included Microsoft - the only time I can say I was in the room with Willy
G. - but that's another story)​.  This work would become the System III
license.

​Xenix, which was V7 based originally, was target for the generic 8086
systems (as well as PDP-11, 68K and Z8000) but the Intel support was
generic so it included the PC.   The bigger problem is that it really
wanted a hard disk, which made its target a 'high end' computer in those
days.  Running on a floppy was sort of possible (besides slow, it tended to
wear out the oxide in the center of the disk where the superblock and
i-list was storied from all r/w - I used to have an 8" floppy sans case
hanging in my office).

But that said, the V7 based license was terrible for a 'small value system'
and I'm not sure Microsoft shipped much against it.   That was the primary
reason they wanted to a new license [Gate's line at the time: 'You guys
don't get it.  The only thing that matters is volume.'].   So until the
System III license, which is what Xenix 2.0 and later shipped, I don't
think Microsoft really had much presence.  But at that time, Microsoft [via
Bob Greenburg a founder and the only one of Gates roommates at Harvard to
graduate btw] was trying to a UNIX porting house, similar to HCR.   That
had been the original vision of Xenix, they would OEM the SW to other firms
that build HW, just like they did for BASIC and were beginning to do for
FORTRAN and COBOL.

As you point out, other firms such as ISC, HCR and Locus appeared on the
scene as more UNIX knowledgeable.   I also think IBM already had placed big
bets on Microsoft for DOS and BASIC, so they wanted to spread the risk a
little which is why ISC got the original UNIX for the PC HW deal.   Once
Microsoft had Xenix stable, IBM was already their customer so selling Xenix
on IBM was a secondary issue.

For a little more on the history front; my former boss, Phill Shevrin, who
would later switch from ISC to Locus where I worked for him, pulled what I
always thought was one of the great tricks of salesmanship in the UNIX
business.  When he was at ISC, he sold for $1.5M a 'port' to the 386 of
System III to each of IBM, AT&T and Intel.   He got paid three times for
same work and got to sell the result as their own product when they were
done.

Also, from a historical standpoint, I hope we do have the ISC 386 code base
- that was the first of the 32-bit linear UNIX ports for the x86 systems.
The other thing that we should try to find is the Phoenix Tech VP/IX code
base and the Locus Merge Code base.   These were the first VM/hypervisors.
They ran in PC/IX 386 and allowed Windows to run under it, long before
VMware existed.

Clem
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20171206/8a5d8fb6/attachment.html>