[Info-vax] VSI licencing policy (again), was: Re: VSI has a new CEO

[John Dallman]

In article <c4f936d9-317d-4ce3-977f-8b492dc81407n at googlegroups.com>,
lawrencedo99 at gmail.com (Lawrence D_Oliveiro) wrote:

> Windows NT was supposedly designed from the beginning to be portable
> across more than just x86, and look at what a failure that was.

It was initially shipped on MIPS, PowerPC and Alpha, and the dropping of
those platforms was because they weren't much used, rather than because
they didn't work. It then shipped on Itanium, which was dropped because
the market preferred x86-64, rather than because it didn't work. x86-64
is now dominant in the market, but ARM64 is shipping and working. 

My employers have shipped Windows NT software on x86, Alpha, Itanium,
x86-64 and ARM64. I did the Itanium, x86-64 and ARM64 ports. Other
engineers did porting work for MIPS and PowerPC, but both of those
platforms were abandoned before being shipped, because their prospective
customers lost interest.

> I suggested on one of VSI's YouTube videos that they rearchitect 
> VMS on top of a Linux kernel. That should simplify the job 
> immensely, since Linux already runs on essentially every major 
> processor architecture still in existence.

That might have been a good idea several years ago, but given they have
it working now, changing would be foolish. The unusual difficulties with
porting VMS seem to have been the need to create BLISS and MACRO-32
compilers, the out-of-date C compiler, and a much smaller team than DEC
had for the Alpha port, or HP for the Itanium port. 

John