[Info-vax] Posts

[bill]

On 5/27/2024 9:16 PM, David Goodwin wrote:
> In article <v337me$92s3$1 at dont-email.me>, arne at vajhoej.dk says...
>>
>> On 5/27/2024 6:07 PM, David Goodwin wrote:
>>> In article <664dfc17$0$705$14726298 at news.sunsite.dk>, arne at vajhoej.dk
>>> says...
>>>>
>>>> On 5/22/2024 9:34 AM, Single Stage to Orbit wrote:
>>>>> On Wed, 2024-05-22 at 12:19 +0000, Simon Clubley wrote:
>>>>>> Another reason could be that many VMS systems have reached the end
>>>>>> of their life and, for various reasons, many have now been replaced
>>>>>> with non-VMS solutions.
>>>>>>
>>>>>> VMS is clearly in a managed decline situation, but the real question
>>>>>> is just how rapid is that decline before there isn't a large enough
>>>>>> userbase left to remain viable ?
>>>>>
>>>>> When it gets to that point I /really/ would like them to put it into
>>>>> the public domain and let us the hackers add drivers and other things
>>>>> to run it bare metal.
>>>>
>>>> That idea has come up numerous times.
>>>>
>>>> Most believe that it is totally impossible.
>>>>
>>>> VSI does not own the rights to all of VMS. VSI has a license
>>>> from HPE for the old parts of VMS and own the right to the
>>>> new parts of VMS that they have added.
>>>>
>>>> The chance of getting HPE to approve open sourcing the stuff
>>>> they own are close to zero. Only cost - no benefits.
>>>
>>> What costs would there be for HPE beyond those already paid as part of
>>> figuring what, if anything, they could sublicense to VSI?
>>
>> All.
>>
>> Whether HP/HPE can give VSI a license to sell VMS binaries similar
>> to how HP/HPE sold them and whether HPE can release the source code
>> as open source under license XYZ are two different questions.
>>
>> And with supposedly 25 million lines, then it will require a significant
>> software engineering and legal effort.
> 
> But HP/HPE didn't just give VSI a license to sell binaries.
> 
> HP/HPE released source code to VSI and allowed VSI to take that code and
> build new things on top of it (VSI OpenVMS).
> 
> I don't see how that is all that different from releasing the source
> code to everyone and allowing everyone to take that code and build new
> things on top of it (Open OpenVMS).

Because VSI, undoubtedy, signed a very strict NDA in order to do this.

> 
> Either way you're distributing the code to someone other than HPE
> employees and you'd have to be certain you had the right to sublicense
> any 3rd party code under your chosen terms before doing that.

They are distributing licensed copies of binaries.  No source contained.
Question, I am sure everyone here remembers the Microfiche Listings.
(Which could not actually be used to actually build VMS.)
Does VSI offer anything similar?

> 
> Given HPE hasn't added anything new since they conducted that review,
> HPEs rights at this point should be known and additional reviews
> shouldn't be necessary.

How much of a revue depends greatly on what the agreement is.  That's
how lawyers make all their money.

> 
> The only potential issue I see is if HPE is having to pay per-license
> royalties to someone else for some 3rd-party thing in OpenVMS. But if
> they're doing that then they'd already know about it and such things
> could be stripped from an open-source release if if were to ever happen,
> just as Sun never open-sourced certain bits of Solaris.
> 
>> I think many underestimate the effort it takes to open source
>> proprietary code. There is a well known example. Sun Java -> OpenJDK.
>> That was a top-priority of the new Sun CEO. But it still took 12
>> months to release 96% of the code. And it took a few more years
>> to get the last 4% replaces with open source.
>>
>> And that was when it was pushed by the CEO. I do not expect
>> that Neri would push for HPE open sourcing VMS the same way.
>>
>> It was not cheap in 2014 either. But back then there was some
>> benefits too - there were customer commitments that HP/HPE could
>> shift over to VSI.
>>
>>> And of course HPE could just wash their hands of OpenVMS and transfer
>>> the copyrights entirely to VSI.
>>
>> Still work with no benefits.
> 
> Presumably if that ever happened it would be a case of VSI buying
> OpenVMS from HPE outright so the benefit would be in the being paid for
> it.

In which case the cost would likely be totally prohibitive.

> 
> The bigger question is if there is any good reason for VSI to open-
> source anything. They're clearly not interested in increasing adoption
> so not a whole lot to gain by open-sourcing it.

Of course, based on the comments we see here over and over again, if
it were doable what is the reason HPe hasn't just thrown the source
out there.  Heck, they throw it to the public domain.  The FSF is not
the only way to open source something.

> 
> Though sometimes this stuff happens for no reason other than people
> within the organisation wanting it to happen. MS-DOS 4.0 wasn't open-
> sourced because Microsoft was after publicity or increasing sales of
> something. My understanding is that it was open-sourced because someone
> asked for it and a few people within Microsoft helped to make it happen.
> The publicity is more of a bonus than a goal.

Probably combined with the fact that the source was probably stolen
years ago anyway.  :-)

bill