[Info-vax] VMS port to x86

[Keith Parris]

On 5/31/2012 6:42 AM, Phillip Helbig---undress to reply wrote:
> Either this clone has the power of VMS or it doesn't.  If it does, then
> why would anyone buy VMS from HP?

Does Linux have the power of AIX, Solaris, and HP-UX?

If so, why do people continue to buy AIX, Solaris and HP-UX?

I think the answer realistically is: For some fraction of the customers, 
Linux is good enough to replace a proprietary UNIX. For the rest, they 
need something special that is provided at this point only in the 
proprietary UNIX (perhaps better stability, better support, or at least 
the ability to run on their existing proprietary hardware) that isn't 
yet available on Linux.

I expect the same to be true of a clone of VMS. Initially it would be 
expected to have only the most-interesting capabilities, but growing 
over time to meet the needs of more and more of the current OpenVMS 
customers. From the testimony of Intel's Kirk Skaugen 
(http://www.ocpol.com/hp-seeks-pretrial-ruling-in-itanium-lawsuit-against-oracle-spun1_1059.html), 
HP can buy Itanium chips through 2022, with the potential to extend the 
contract, but what about after that?

Looked at another way: The Linux ecosystem is not based on selling Linux 
itself (purchase price) like the Windows ecosystem, but on selling 
support services layered on top of the free open source Linux code. HP 
happily sells hardware which runs Linux (or Windows), although those 
both arguably compete with HP-UX sales. HP would likely do the same with 
a VMS clone.

> (This leaves aside the question whether a VMS clone would be legal.)

To be legal, you need to avoid infringing copyrights, patents, and trade 
secrets. Clean-room techniques have proven successful. Linux proved to 
be legal, despite SCO's best efforts. Even without a clean-room start, 
OpenBSD eventually became legal, after replacing some bits of offending 
UNIX code.

Most of the basic VMS patents (e.g. VAXclusters, SCS, MSCP) are now 
expired. There are still likely some in-force patents on recent 
developments like host-based mini-merge and such. So for those you 
either forgo the feature or get dispensation from HP to use the feature 
with a non-royalty license (and HP makes money off sales of hardware 
running the VMS clone).

Given the direction and emphasis of Project Odyssey, do you think HP 
would be likely to sue a group of individuals that provide a direction 
into the future for VMS customers on x86 hardware?

On 5/31/2012 6:42 AM, Phillip Helbig---undress to reply wrote:
 > In article<slrnjseja2.ul.jkb at rayleigh.systella.fr>, JKB
 > <jkb at koenigsberg.invalid>  writes:
 >> 	Ten years ago, I have started with some other VMS addicts this
 >> 	kind of project.
 >
 > But doesn't this illustrate the problem?  10 years is a long time
 > in IT, and there is no usable VMS clone to come out of this.

Up 'til now, most people haven't perceived the need as urgent enough. 
Even now, with at least 10 years of Itanium hardware availability ahead, 
the need may not yet be perceived as urgent enough in the customer base; 
we'll have to see.

 > While I'm willing to believe that a Finnish student could write a unix
 > kernel in a few weeks, I fail to see how even a large group---say 100
 > people working full time---could even hope to come up with something
 > which looks like VMS. Think of all the resources which went into
 > creating VMS in the first place.

GNU started when AT&T purchased a stake in Sun and started restricting 
access to UNIX. There was concern that access to UNIX would be lost. I'm 
confident the VMS community can do as well as the GNU/Linux folks did, 
if driven with the same motivation.