[Info-vax] Vaxes shutting off this week

[Glen Herrmannsfeldt]

JF Mezei wrote:
> Michael Moroney wrote:

>>A big difference is that the PDP and the VAX ran their courses and were
>>discontinued at an appropiate time.  Alpha was killed when it was pretty
>>much the fastest processor. 
(snip)

> VAX was never really allowed to break its barrier because
> Alpha came too quickly to market.

I believe that VAX made some decisions wrong in the beginning,
which were bad for it later.  Consider that S/360 has been extended
twice and and is still a viable architecture.  S/360 (and now
z/Architecture) has only three different instruction lengths.
With that is isn't so hard to do instruction decode for multiple
instructions in a cycle. You can deduce the instruction length
from the first byte in all cases.

With VAX and address mode bytes it is horribly complicated
to decode instruction boundaries.  VAX works well for a
microprogrammed serial instruction processor, reading bytes
and operating on them.

VAX was well designed for assembly language programming just
at the time that more and more programming, including systems
programming, was done in HLLs.

> So, for a 25 year timeline, Alpha was the right choice beause it offered
>  growth for the whole period. But for a 10 year history, it would have
> been cheaper to just upgrade VAX until 25-JUN-2001 and forget about the
> port to ALpha and the abandonment of so much software.

-- glen