Zane Healy wrote:
On Nov 12,
2020, at 10:39 AM, Peter Coghlan via cctalk <cctalk at classiccmp.org> wrote:
There's something icky about VMS running on x86. I can't see the prospect
being attractive to either the VMS people or the x86 people.
Do you have any valid data to base that statement on?
None whatsoever, only my own observations and opinions.
Quite a few people on the x86 side that I?ve talked to are interested
by the prospect, and from what I?ve seen, it?s being well received on
the VMS side. I know of one company that is interested in seeing if they
should add support in their product for it.
Over the years, I've seen all sorts of people enthusing about how great
they think VMS would be, if only it:
- could use commodity SCSI disks
- could be orders of magnitude faster
- could do networking with this, that or the other
- could be more compatible with Unix
- could be more compatible with Windows
- could have a goofy graphical user interface
- could have native support for flavour of the month programming language x
- could support comodity ISA/EISA/PCI/USB/whatever flavour of the month I/O
- could be run on commodity hardware used for other operating systems
- could be used for free by students and hobbyists
- etc
All of these things and more were delivered in spades or were there from
the beginning and yet it never seems to be enough. Some people are just
never satisfied, not to mention it never seems to dawn on them just how
useful VMS clustering is compared to anything else they have seen.
Mark Daniel already has WASD up and running on OpenVMS/x86.
I don't want to belittle Mark's achievement here but if this x86 port is
going to be in any way worthwhile, it shouldn't be a big deal to build a
well behaved, long established VMS application on it.
There have already been 6 customer releases of the x86 version, with a 7th
due in about a month, in fact the V9.0-F release looks to be pretty
significant
(I believe it?s where cluster support will start showing up).
It doesn't even have cluster support yet???
The problem with the x86 port is when you have software that only runs on
older versions or architectures. This is also a big problem for the
Itanium port. For example I keep a system running VAX/VMS 5.5-2 for just
this reason, and there is a ton of software that is VAX only, or at best
Alpha only.
I've had plenty of success doing binary translation of VAX software to Alpha
when source was not available. I never had any interest in the Itanium
platform so I don't know if the same could be done there.
One thing that is interesting about the x86 version is that people will be
able to easily get their feet wet with a modern version of the OS.
Why can't people do that now using Alpha emulation running on their platform
of choice?
I?m anxiously awaiting VMware and Hobbyist support.
If VMware was worth it's salt, it would be transparent to whatever is running
under it, not needing specific support in whatever is running under it.
Anything that makes it easier to get OpenVMS into the hands of hobbyists
and students is a good thing.
I've been hearing that for 30 years or more and VMS doesn't seem to have
taken over the world yet.
Regards,
Peter Coghlan
Zane