Linux vs Windows, obscure security features (was: Quarantining an account...)

[Jon 'maddog' Hall]

Ben,

>From an admittedly faulty and ever-aging memory of events:

>   And Cutler moved to Microsoft because DEC just wanted to
> maintain/extend VMS, while Cutler wanted to write a new OS ("MICA")
> for the new hardware architecture ("PRISM") that was being designed.
> Microsoft needed a better OS (where "better" included "not part-owned
> by IBM"), Cutler wanted to continue MICA... and thus "OS/2 NT" was
> born.  It was originally going to be the 3.0 release of OS/2.  Then
> the IBM and MSFT alliance fell apart completely, and it became
> "Windows NT".
> 
Cutler wanted to leave Massachusetts and live in Washington State a long
time before that.  KO wanted to keep him on board, so allowed him to set
up an advanced development facility in Belleview, overlooking the
Olympic Mountains, which Cutler could see perhaps three days a year)
funded by Digital.

Dave was well on his way to developing an operating system based on
Microkernel technology that could run both a VMS personality and a Unix
personality at the same time as well as developing the Prism hardware
architecture.

However, because this was an extensive and expensive operation, Dave's
budget was often reduced, and the project sometimes was threatened with
closure, but eventually was refunded and continued on.

The other problem was that Dave was of the opinion that MICA had to be
"exactly" VMS compatible, whereas Dave could change Unix when there were
some conflicts between the two OS designs, or where he could "make it
better".  I know this because I interviewed for a job as product manager
of the Unix side, and spent a futile twenty minutes trying to convince
him he was wrong...not a very good thing to do in a job interview.

Meantime Digital was being pulverized by Sun and Sparc workstations, and
an engineer from the east coast, Rob Rodriguez had the idea (inspired by
John Hennessy) of porting Ultrix (a VAX-based Digital extension of BSD
4.1) to a MIPs-based little-endian workstation based on the VAXstation
3100 motherboard (and subsequently called the DECstation 3100) giving a
several-times performance improvement over then-current VAX
architectures.  This idea was seized by Armando Stettner, floated to the
west coast workstation manager, Joe DiNucci, who then took it to Ken
Olsen who immediately funded it.

This project (called the PMAX project) was done in complete secrecy,
with a minimal engineering team.  I still have my "PMAX" T-shirt some
place.

On the eve of the announcement of the DecStation 3100, Ken Olsen called
a Board of Directors meeting, and invited Dave to fly east and talk
about the status of his work.

After Dave put all of his charts, drawings, plans and schedules up in
front of the Board, KO asked him how fast the new system would be, and
would it be any faster than the prototype Decstation that KO rolled out
in front of him?

Dave left the meeting, flew back to Belleview, and a short time later
(within two weeks, I believe) left DEC and joined Microsoft.  Some of
his engineers went with him.  About six months later Microsoft announced
a brand new OS called Windows NT, for "Windows New Technology", that had
been developed by Dave Cutler and his staff.

In cleaning up the remains of the advanced development lab, it is
rumored that a silicon wafer with some CPU prototypes were found, and
when tested were a very fast RISC processor originally called the
"E-VAX" (for "Extended VAX" but later developed into the Alpha series.

>  Reportedly, the NT kernel and VMS share a number of architectural
>similarities.  I read a 2-page technical analysis once; most of it was
>over my head but it sure seemed like there was something to it.  I've
>been told NT was so similar DEC threatened MSFT with legal action, and
>MSFT settled out-of-court; one consequence was that NT was maintained
>on the Alpha for longer than had MSFT wanted.  Or something like that.

Whether that OS had any Digital Proprietary IP in it is up to the
lawyers to determine, but it is interesting to think about someone (even
as brilliant as Dave is) who could write an entire OS in such a short
time without using some existing IP.

So now you know "the rest of the story" (or at least as it was
remembered by me).

md