[Info-vax] VMS port to x86

[David Froble]

JKB wrote:
> Le Thu, 24 May 2012 02:59:48 -0400,
> JF Mezei <jfmezei.spamnot at vaxination.ca> écrivait :
>> Neil Rieck wrote:
>>
>>> If I were HP, I'd have the skunk-works in India working on an x86 port
>>> right now.
>> Thyey tried this a few years ago but those skunkworks which also
>> involved HP-UX were killed because moving HP=UX to x86 would be too much
>> of an effort.
>>
>> With regards to running VMS in emulation. VAX-VMS is over 10 years old
>> at 7.3. Not developped anymore. And Alpha/IA64 is barely being
>> maintained/developped anymore. About the only commitment from HP is to
>> make it run on the upcoming Poulson and Kittson platforms.
>>
>> So even if you can continue to run VMS on emulated environments, you are
>> still limited to whatever VMS version and software are available.
> 
> 	Right. I think it's really time to know if we want to assist to VMS
> 	death or if we want to do somethink to keep VMS alive. Even if VMS
> 	is a fabulous OS, its design has to be changed. A long time ago, Dec
> 	has started to port VMS on a microkernel (VMS-O-Mach) and I'm sure
> 	that if VMS has today still future, it has to be rewritten over a
> 	microkernel. It's time to drop Bliss and some strange languages to
> 	write VMS as servers (written in C, ADA or all language you want)
> 	on top of a real microkernel (not Mach, but L4 for example).
> 
> 	But VMS is dying also as porting Unix application to VMS is not
> 	simple. If you want to propose a lot of softwares, you cannot ignore
> 	Unix and VMS should have a real POSIX/SysV libc like newlib.
> 
> 	Regards,
> 
> 	JKB
> 

I do not share this perspective.  Not saying which if either is a better perspective.  As 
for Bliss, I have little experience working with it, but it seems to be adequate.  If 
there is a Bliss compiler, what is the problem?

The language I'd prefer to drop is C, but that's just personal preference, which is what I 
consider the above suggestion.

VMS does not have to look and feel like Unix in order to be successful.  If that were the 
goal, then why not just use Unix?

What VMS needs is a viable hardware platform, good implementations of current 
capabilities, and continued development of new capabilities as they come into use.