[Info-vax] Python on VMS
Bill Gunshannon
bill.gunshannon at gmail.com
Wed Jan 16 17:17:46 EST 2019
On 1/16/19 4:59 PM, Dave Froble wrote:
> On 1/16/2019 1:17 PM, Phillip Helbig (undress to reply) wrote:
>> In article <q1nds2$hic$1 at gioia.aioe.org>, =?UTF-8?Q?Arne_Vajh=c3=b8j?=
>> <arne at vajhoej.dk> writes:
>>
>>> Would it be fair to say that the problems for the port is
>>> not fundamental (*nix'isms that simply can't be implemented
>>> on VMS) but a resource problem (if there were enough VMS
>>> enthusiasts working on those libraries then they would be
>>> ported)?
>>
>> It's obviously a resource problem because there is a theorem which says
>> that any Turing machine can implement another. :-)
>>
>
> That could have several meanings.
>
> Implement Unix on VMS? Why not just run Unix?
Or implement a workable POSIX interface on VMS.
>
> For software to be portable, the people implementing it need to design
> for such. From the little I've read, some of those developing for Unix
> don't seem to care if their product can be used on VMS, and sometimes
> use capabilities that will never work on VMS.
Two errors in this statement. "some of those developing for Unix
don't seem to care if their product can be used on VMS". Make that
"all". And the second is you made the assumption those developers
have ever heard of VMS, much less seen, used, or developed for.
>
> Sometimes some things just aren't feasible.
I think the Software Tools Virtual Operating System proved that
wrong almost 40 years ago. Too bad that project lost it's momentum
and we had to wait for POSIX to start all over from scratch with a
pre-conceived notion that it wasn't going to work.
>
> I have to ask, if a product is understood, why not just implement it on
> VMS, using VMS capabilities? Might be easier. But then, it would not
> be portable back to whatever. So what, from a VMS centric perspective?
Personally, I agree with you there. It would likely make keeping
up to date more work, but I think it would provide better products
and, in the long run, require less real effort.
I have been setting up a whole world of systems for some of the
stuff I want to play with again. I am utterly amazed (well, more
dismayed) at what a morass most of the Unixes have become since
I was doing this on a daily basis. Building stuff from "PORTS"
trees and watching hundreds if not thousands of warnings go by
during the compile. Some of them downright scary. And the
current autoconfigure that I believe people are trying to
duplicate on VMS is a nightmare. Better to just take the source
for a desired product and put it on VMS as a VMS product. You
can still look at upgrades using DIFFS but stop wasting time
trying to duplicate a convoluted and frightening configuration
scheme.
bill
More information about the Info-vax
mailing list