[Info-vax] Kittson question

[johnwallace4 at yahoo.co.uk]

On Monday, 5 January 2015 22:13:16 UTC, Stephen Hoffman  wrote:
> On 2015-01-05 21:49:22 +0000, Stanley F. Quayle said:
> 
> > On Sun, 2015-01-04 at 09:33 -0500, Scott Dorsey via Info-vax wrote:
> >> I think this is much more an issue for ELN and Ultrix
> >> users than VMS, but I don't think there are any of those left.
> > 
> > Without naming sites, let me just say that the defense of our nation 
> > rests on VAXELN.
> 
> Um, so?
> 
> Cuba still uses a number of 1950s-era Detroit cars.
> 
> The folks involved will undoubtedly be replacing both the old cars and 
> the old VAX ELN stuff, when the opportunity next presents
> 
> Both the old cars and the old VAX ELN gear do solve the immediate 
> problems, and both can certainly be kept going, but neither is 
> particularly effective, efficient nor easy to get parts for.  Not by 
> present-day standards.  This stuff works, but it is just not what folks 
> would roll out for replacements and for new deployments.
> 
> Pointing to some ancient PDP-11 boxes running nuclear reactors has the 
> same basic problem.  There are other constraints.   Outside of those 
> constraints, none of this gear is where the market is now, and none of 
> this gear is where the market is headed.
> 
> 
> 
> -- 
> Pure Personal Opinion | HoffmanLabs LLC

Back in the day, DEC E+RT HQ decided not to develop a VAXELN successor
on Alpha, and instead decided that it would be sensible if VAXELN
customers were offered the opportunity to migrate their VAXELN
applications to a VAXELN emulation layer on top of a commercial
RTOS, VxWorks, targeting either M68K or Alpha. VxWorks comes from
Wind River Systems, now owned by Intel.

The emulation layer was largely pointless as the two OSes are so
dissimilar that a move would typically be a redesign rather than a
minor porting exercise (and that's without considering things like
the requirement to move off VMS as development host to UNIX or
Windows as development host).

No one I knew of chose to make the recommended move.

Nothing else out there was like VAXELN at the time, especially for
people building non-trivial distributed RT applications. 

There was an E+RT HQ vision of a "next generation" distributed RT OS,
building (incompatibly) on some of the lessons learned from VAXELN
and subsequent developments. Afaik, nothing ever came of it inside
DEC/CPQ/HP, and when E+RT was sold off to SMART (and then Motorola
and then Emerson and ...) it seemed to vanish.

I'm told that VAXELN had a rather nice implementation of the
"channels and pools" concepts from Tony Hoare's "Communicating
Sequential Processes", which is widely regarded as a classic work
in the RT community (it's now legitimately freely downloadable at
www.usingcsp.com, if anyone's interested). 

There are lots of embedded and/or real-time OSes and toolsets to
choose from today. I've not seen anything as 'elegant' as VAXELN,
though as usual there are plenty of more widely used possibilities.