Let's open source VMS!

[Single Stage to Orbit]

With the news that HP is to take VMS out the back and put a bullet
through its brains, can we hope that HP will open source the whole thing
so that people can keep using it?
--
Tactical Nuclear Kittens

[Arne Vajhøj]

On 6/11/2013 6:21 PM, Single Stage to Orbit wrote:
> With the news that HP is to take VMS out the back and put a bullet
> through its brains, can we hope that HP will open source the whole thing
> so that people can keep using it?

You can hope.

But it does not seem that likely.

Arne

[dgso...@gmail.com]

They don't have to open-source the *whole* thing. Just open-sourcing the bits they hold the copyrights on (which I expect is the majority of the core system) may be enough. Replacements can probably be written for whatever HP can't release (as happened with OpenJDK, Illumos, Haiku, etc).

I'll admit seeing an Open OpenVMS seems unlikely but HP may not entirely object to the idea - they did open-source AdvFS a while back after all...

[JF Mezei]

On 13-06-11 20:28, Arne Vajhøj wrote:

> You can hope.
> But it does not seem that likely.

HP bought Palm and a short time later abandonned its investment and
announced it would open source the abandonned OS.

This was very proprietary OS and I suspect encumbered with a lot of
stuff. And HP never made any money from it, so justifying the spending
to determine what could and couldn,T be made open source was even less
justified than similar process for VMS which made HP lots of money over
the years.

[Bill Gunshannon]

In article <5gkk8a-...@nntp.local.net>,

Single Stage to Orbit <alex....@munted.eu> writes:
> With the news that HP is to take VMS out the back and put a bullet
> through its brains, can we hope that HP will open source the whole thing

> so that people can keep using it?=20

Not a snowball's chance in hell.

bill

--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bill...@cs.scranton.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>

[dgso...@gmail.com]

On Wednesday, June 12, 2013 2:44:37 PM UTC+12, Bill Gunshannon wrote:
> In article <5gkk8a-...@nntp.local.net>,
>
> Single Stage to Orbit <alex....@munted.eu> writes:
>
> > With the news that HP is to take VMS out the back and put a bullet
>
> > through its brains, can we hope that HP will open source the whole thing
>
> > so that people can keep using it?=20
>
>
>
> Not a snowball's chance in hell.

Is there any particular legal or technical reason why HP couldn't open-source it? Assuming they haven't sold it they presumably still hold copyrights for most of the core platform. If approval was given all someone would really need to do is slap a new license header and copyright statement at the top of each file HP hold copyrights for and dump it on an FTP server somewhere.

It may just take some motivated people with contacts in HP to make something happen. Considering the alternative I'd think it would certainly be worth a try...

Of course if everyone within the community is content to see OpenVMS ultimately go extinct then without a doubt that is what will eventually happen.

[JKB]

Le Tue, 11 Jun 2013 23:21:24 +0100,
Single Stage to Orbit <alex....@munted.eu> écrivait :

> With the news that HP is to take VMS out the back and put a bullet
> through its brains, can we hope that HP will open source the whole thing
> so that people can keep using it?

You can hope. But I think that rewrite an OpenVMS clone from scratch
will be better and more efficient than convert (even if real OpenVMS
will be opensourced) OpenVMS to a new programming language to be portable.
Don't forget that OpenVMS runs only on IA64 and I'm not sure that
IA64 has a future.

JKB

--
http://www.freevms.net

[VAXman-]

I could swear that I was running OpenVMS on VAX and on Alpha.

--
VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)ORG

Well I speak to machines with the voice of humanity.

[JKB]

Le Wed, 12 Jun 2013 11:37:35 GMT,
VAXman- @SendSpamHere.ORG <VAX...@SendSpamHere.ORG> écrivait :

> In article <slrnkrgb8...@rayleigh.systella.fr>, JKB <j...@koenigsberg.invalid> writes:
>>Le Tue, 11 Jun 2013 23:21:24 +0100,
>>Single Stage to Orbit <alex....@munted.eu> écrivait :
>>> With the news that HP is to take VMS out the back and put a bullet
>>> through its brains, can we hope that HP will open source the whole thing
>>> so that people can keep using it?
>>
>> You can hope. But I think that rewrite an OpenVMS clone from scratch
>> will be better and more efficient than convert (even if real OpenVMS
>> will be opensourced) OpenVMS to a new programming language to be portable.
>> Don't forget that OpenVMS runs only on IA64 and I'm not sure that
>> IA64 has a future.
>
> I could swear that I was running OpenVMS on VAX and on Alpha.

Sure. But I think that today, if you want run OpenVMS on new
hardware, you have to buy an IA64. You can try to buy nuVAX or an
AXP emulator, but it's not a solution.

--
http://www.freevms.net

[Bob Gezelter]

JKB,

There is vanishingly little of the OpenVMS source base that is architecture-specific to IA64. The vastly overwhelming majority of the "assembler" level code is written in MACRO-32, which is compiled for both Alpha and IA64.

Having been down both paths (re-writing vs modifying compilers) in the past, I can assure you that it is almost without a doubt more effective to do a port by redoing the translator than to do a recode.

In this case, a port to x64 would require a new backend for the MACRO32 compiler, the BLISS compiler, and one or two others. Work, but not as challenging as a complete re-write.

- Bob Gezelter, http://www.rlgsc.com

[JKB]

Le Wed, 12 Jun 2013 06:12:58 -0700 (PDT),
Bob Gezelter <geze...@rlgsc.com> écrivait :

> On Wednesday, June 12, 2013 7:42:00 AM UTC-4, JKB wrote:
>> Le Wed, 12 Jun 2013 11:37:35 GMT,
>>
>> VAXman- @SendSpamHere.ORG <VAX...@SendSpamHere.ORG> écrivait :
>>
>> > In article <slrnkrgb8...@rayleigh.systella.fr>, JKB <j...@koenigsberg.invalid> writes:
>>
>> >>Le Tue, 11 Jun 2013 23:21:24 +0100,
>>
>> >>Single Stage to Orbit <alex....@munted.eu> écrivait :
>>
>> >>> With the news that HP is to take VMS out the back and put a bullet
>>
>> >>> through its brains, can we hope that HP will open source the whole thing
>>
>> >>> so that people can keep using it?
>>
>> >>
>>
>> >> You can hope. But I think that rewrite an OpenVMS clone from scratch
>>
>> >> will be better and more efficient than convert (even if real OpenVMS
>>
>> >> will be opensourced) OpenVMS to a new programming language to be portable.
>>
>> >> Don't forget that OpenVMS runs only on IA64 and I'm not sure that
>>
>> >> IA64 has a future.
>>
>> >
>>
>> > I could swear that I was running OpenVMS on VAX and on Alpha.
>>
>>
>>
>> Sure. But I think that today, if you want run OpenVMS on new
>>
>> hardware, you have to buy an IA64. You can try to buy nuVAX or an
>>
>> AXP emulator, but it's not a solution.
>>
>>
>>
>> --
>>
>> http://www.freevms.net
>
> JKB,
>
> There is vanishingly little of the OpenVMS source base that is architecture-specific to IA64. The vastly overwhelming majority of the "assembler" level code is written in MACRO-32, which is compiled for both Alpha and IA64.

I know. I have the listing of 8.2.

> Having been down both paths (re-writing vs modifying compilers) in the past, I can assure you that it is almost without a doubt more effective to do a port by redoing the translator than to do a recode.
>
> In this case, a port to x64 would require a new backend for the MACRO32 compiler, the BLISS compiler, and one or two others. Work, but not as challenging as a complete re-write.
>
> - Bob Gezelter, http://www.rlgsc.com

Sure. But you suppose that HP will opensource OpenVMS. If you write
a clone, you can start to write this clone today. For information,
there is a Bliss compiler written as gcc front end.

JKB

--
Si votre demande me parvient sur carte perforée, je titiouaillerai très
volontiers une réponse...
=> http://grincheux.de-charybde-en-scylla.fr

[Bill Gunshannon]

In article <d520f739-10eb-4dec...@googlegroups.com>,

dgso...@gmail.com writes:
> On Wednesday, June 12, 2013 2:44:37 PM UTC+12, Bill Gunshannon wrote:
>> In article <5gkk8a-...@nntp.local.net>,
>>
>> Single Stage to Orbit <alex....@munted.eu> writes:
>>
>> > With the news that HP is to take VMS out the back and put a bullet
>>
>> > through its brains, can we hope that HP will open source the whole thing
>>
>> > so that people can keep using it?=20
>>
>>
>>
>> Not a snowball's chance in hell.
> Is there any particular legal or technical reason why HP couldn't
> open-source it?

It would take an army of engineers and lawyers to go thru every single
line of code to determine that they actually have the right to do it.
Someone has to pay for that. HP is very unlikely to be willing to
accept that cost as they would be getting nothing in return and it
is just as easy to just destroy everything and let it die.

> Assuming they haven't sold it they presumably still
> hold copyrights for most of the core platform.

An assumption that would take that army of engineers and lawyers to
determine.

> If approval was given
> all someone would really need to do is slap a new license header and
> copyright statement at the top of each file HP hold copyrights for
> and dump it on an FTP server somewhere.

And even that would cost quite a bit and again, what would HP be getting
out of it to justify the expense? Good will? Hardly. Most of the IT
world is unaware of VMS and probably couldn't care less. That leaves
a handful of C.O.V denizens who want to continue running it in their
basement. hardly a justification for the cost or the potential liability.

> It may just take some motivated people with contacts in HP to make
> something happen. Considering the alternative I'd think it would
> certainly be worth a try...

Worth it to who? You? HP?

> Of course if everyone within the community is content to see OpenVMS
> ultimately go extinct then without a doubt that is what will eventually
> happen.

One need only look at the fate of other similar products. Once again I
have to bring up the PDP-11 OSes. Probably less than a tenth the number
of lines of source code and all attempts to pry it from the hands of its
owner since its demise have been fruitless. And guess who the top of
the IP foodchain for these OSes is. :-)

[David Froble]

I think that JF has been outdone. We have a new #1 ney-sayer ....

VMS is more than just a bunch of hobbyists. But you know that. You're
just having fun. why don't you go change the oil in the MG ??

[Simon Clubley]

On 2013-06-12, JKB <j...@koenigsberg.invalid> wrote:
> Le Wed, 12 Jun 2013 06:12:58 -0700 (PDT),
> Bob Gezelter <geze...@rlgsc.com> écrivait :
>>

>> JKB,
>>
>> There is vanishingly little of the OpenVMS source base that is architecture-specific to IA64. The vastly overwhelming majority of the "assembler" level code is written in MACRO-32, which is compiled for both Alpha and IA64.
>
> I know. I have the listing of 8.2.
>
>> Having been down both paths (re-writing vs modifying compilers) in the past, I can assure you that it is almost without a doubt more effective to do a port by redoing the translator than to do a recode.
>>
>> In this case, a port to x64 would require a new backend for the MACRO32 compiler, the BLISS compiler, and one or two others. Work, but not as challenging as a complete re-write.
>>
>> - Bob Gezelter, http://www.rlgsc.com
>
> Sure. But you suppose that HP will opensource OpenVMS. If you write
> a clone, you can start to write this clone today. For information,
> there is a Bliss compiler written as gcc front end.
>

There's also the fact that VMS requires much more hardware support than
other operating systems do. Writing a new backend isn't going to do anything
about that. :-)

However (IIRC) you have said in the past that you have been able to use
the microkernel design approach to emulate the multiple protection rings
within your FreeVMS microkernel design.

BTW, what is the location of the Bliss frontend ? I could only find
messages talking about it on the gcc mailing list, but could not find
the kit itself. (I'm not about to start using it in projects, but it
would be interesting to look at the kit.)

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world

[John Wallace]

On Jun 11, 11:21 pm, Single Stage to Orbit <alex.bu...@munted.eu>
wrote:

I'm told it's now the 21st century. Allegedly Web 2.0 and the like
provide mechanisms to find out how interesting a project like this
would be, and whether there's any real potential backing out there.

Who knows enough about crowdsourcing (kickstarter etc) to go about
setting something up?

Who thinks enough people would be willing to commit (donate?) a
worthwhile sum of money to get something going far enough to get HP's
attention?

I know nothing about this stuff, I don't even know whether I think
this might be a good idea let alone whether I think it stands any
chance of success, but in theory it ought to be a possibility in the
age of LinkedIn communities and crowdsourcing and other such
trendiness.

If nothing else, it might help answer the "worth it? to who?"
question. It still leaves a few challenges left to overcome (e.g. the
compiler backends Bob mentions, and maybe the copyright issues
routinely mentioned, though for those there still aren't any
functional showstoppers I'm aware of).

[JF Mezei]

On 13-06-12 10:26, Bill Gunshannon wrote:

> It would take an army of engineers and lawyers to go thru every single
> line of code to determine that they actually have the right to do it.

I have to assume that every module has copyright notices on it and if it
contains other people's code, it would show in the module header. So
they don't need to parse every single line of code.

> Someone has to pay for that. HP is very unlikely to be willing to
> accept that cost as they would be getting nothing in return and it
> is just as easy to just destroy everything and let it die.

Then why did HP decide to go through that trouble for its PalmOS/WebOS ?

> out of it to justify the expense? Good will? Hardly. Most of the IT
> world is unaware of VMS and probably couldn't care less.

You may not see a port of VMS as a whole to another platform. But you
might see a port of say TPU to other platforms, concepts of the DLM and
clustering used on other platforms etc. And if it runs on x86, it could
help HP make techical support sales.

Imagine logical names, with the LMF$SYSCLUSTER concepts allowing logical
names to exists across Linux nodes.

And if HP were even to do the ports of those to Linux, it might help the
remaining VMS custoemr base off VMS faster and start buying x86 servers
from HP along with Linux support.

Of course, this requires a corporation that has more than beancounters
at the top and which has vision and a sense of responsability towards it
very profitable enterprise customers.

Remember that it is HP that insisted VMS customers move to that IA64
thing and it is HP which is pulling the plug on it and VMS.

[Stephen Hoffman]

On 2013-06-12 17:13:50 +0000, Simon Clubley said:

> BTW, what is the location of the Bliss frontend ? I could only find
> messages talking about it on the gcc mailing list, but could not find
> the kit itself. (I'm not about to start using it in projects, but it
> would be interesting to look at the kit.)

Scroll down to the bottom of
<http://en.wikipedia.org/wiki/BLISS_(programming_language)> for what
looks to be the download link.


--
Pure Personal Opinion | HoffmanLabs LLC

[Stephen Hoffman]

On 2013-06-12 17:50:34 +0000, JF Mezei said:

> Imagine logical names, with the LMF$SYSCLUSTER concepts allowing
> logical names to exists across Linux nodes.

Cluster-wide logical names? For what purpose?

Stinky-bad-clumsy configuration data storage? There are better ways to
do that.

Depending on the goal, there are solutions built on mDNS, LDAP, DHCP, etc.

For one approach on OS X, issue /man dns-sd/ and have a look around.

Working on OS X, I have yet to have any "I really need a logical name
here" moment. Maybe that'll change.

{Um, LMF...? I'll assume that was a typo and you're not looking for
distributed licensing, though there are solutions there, too.}

[Bill Gunshannon]

In article <kpaa7u$f7s$1...@dont-email.me>,

What's the difference between a nay-sayer and a realist? I spent many
years fighting to keep VMS in the university environment. I went head-
to-head with their management trying to explain why their "education"
program was not going to fly. We all know where that ended. (I just
asked at the datacenter today and all VMS is now gone from here, not
just from the academic side.) I have been involved in a number of
other cases where a desire exosted to save some piece of defunct
software. None of them were ever succesful. In at least two cases
I was directly involved in the current holder of the IP openly admited
that they deliberately destroyed all copies of the sources.

Go ahead, get your hopes up. Sometimes it takes getting knocked down
real hard to get a point across.

> VMS is more than just a bunch of hobbyists. But you know that. You're
> just having fun. why don't you go change the oil in the MG ??

Is it? At this point, anyone who is not just a hobbyist who is betting
their business on VMS is a fool. It's not like the writting hasn't been
on the wall for a long time.

As for the MG, fuel pump is on the way and I expect to be driving it
in about two weeks. I would expect everyone here to be an MG owner.
We all like antique systems and we are all gluttons for punishment.

[Bill Gunshannon]

In article <51b8b4ec$0$57828$c3e8da3$eb76...@news.astraweb.com>,

JF Mezei <jfmezei...@vaxination.ca> writes:
> On 13-06-12 10:26, Bill Gunshannon wrote:
>
>> It would take an army of engineers and lawyers to go thru every single
>> line of code to determine that they actually have the right to do it.
>
> I have to assume that every module has copyright notices on it and if it
> contains other people's code, it would show in the module header. So
> they don't need to parse every single line of code.
>

You can assume that. Your corporate lawyers will tell you assume
nothing. Does no one remember the fiasco when AT&T removed all of
the BSD Copyright messages from source they incorporated because
some low level data entry type was told that nothing but AT&T
Copyrights should be in any of their source files?

>
>
>
>> Someone has to pay for that. HP is very unlikely to be willing to
>> accept that cost as they would be getting nothing in return and it
>> is just as easy to just destroy everything and let it die.
>
> Then why did HP decide to go through that trouble for its PalmOS/WebOS ?

Don't know, but PalmOS wasn't as old and hadn't been touched by nearly
as many people. And probably wasn't nearly as big.

>
>> out of it to justify the expense? Good will? Hardly. Most of the IT
>> world is unaware of VMS and probably couldn't care less.
>
> You may not see a port of VMS as a whole to another platform. But you
> might see a port of say TPU to other platforms, concepts of the DLM and
> clustering used on other platforms etc. And if it runs on x86, it could
> help HP make techical support sales.

If it isn't a port of VMS why wold anyone here care? I thought VMS
was what everyone wanted? TPU? There are hundreds of editors out
there. TPU is just another. DLM? There has been a freely available
DLM for at least a decade. Pieces of VMS are of little if any value.
It is VMS as a system that has value and that everyone here wanted to
see (or so I thought!)

>
> Imagine logical names, with the LMF$SYSCLUSTER concepts allowing logical
> names to exists across Linux nodes.

That's a VMSism. There are other ways of doing things than the VMS
way. If you are running VMS you do it the VMS way. If your not you
use the local way of accomplishing the same task.

>
> And if HP were even to do the ports of those to Linux, it might help the
> remaining VMS custoemr base off VMS faster and start buying x86 servers
> from HP along with Linux support.

Funny, everyone here seems to be saying that if HP kills VMS they
will never deal with HP again. So, which way is it? Run VMS or
be happy with HP Linux boxes?

>
> Of course, this requires a corporation that has more than beancounters
> at the top

If they had real beancounters at the top Itanic and VMS would have been
gone a long time ago.

> and which has vision and a sense of responsability towards it
> very profitable enterprise customers.

Read the letter again. BCS is bleeding red. It is anything but
profitable. Ink is profitable.

>
> Remember that it is HP that insisted VMS customers move to that IA64
> thing and it is HP which is pulling the plug on it and VMS.

What's your point? People have been predicting this for ages. Anyone
who did not see this coming must have been blind. The only curious
thing is why it took so long....

[David Froble]

Bill Gunshannon wrote:
> In article <kpaa7u$f7s$1...@dont-email.me>,
> David Froble <da...@tsoft-inc.com> writes:
>> Bill Gunshannon wrote:
>>> In article <d520f739-10eb-4dec...@googlegroups.com>,
>>> dgso...@gmail.com writes:
>>>> On Wednesday, June 12, 2013 2:44:37 PM UTC+12, Bill Gunshannon wrote:
>>>>> In article <5gkk8a-...@nntp.local.net>,
>>>>>
>>>>> Single Stage to Orbit <alex....@munted.eu> writes:
>>>>>
>>>>>> With the news that HP is to take VMS out the back and put a bullet
>>>>>> through its brains, can we hope that HP will open source the whole thing
>>>>>> so that people can keep using it?=20
>>>>>
>>>>> Not a snowball's chance in hell.
>>>> Is there any particular legal or technical reason why HP couldn't
>>>> open-source it?
>>> It would take an army of engineers and lawyers to go thru every single
>>> line of code to determine that they actually have the right to do it.
>>> Someone has to pay for that. HP is very unlikely to be willing to
>>> accept that cost as they would be getting nothing in return and it
>>> is just as easy to just destroy everything and let it die.

I question this line of argument.

If there are royalties being paid, HP surely knows to whom and why.

I'd think that HP has the right to sell, or give away, anything in VMS.
Now, the rights of whoever gets the software might be another issue.
But it would be their issue, not HPs. So in such a case there would not
be this assumed cost to HP.

>>>> Assuming they haven't sold it they presumably still
>>>> hold copyrights for most of the core platform.
>>> An assumption that would take that army of engineers and lawyers to
>>> determine.

Why? If someone has a problem with the users of the open sourced VMS,
they can inform said users of their rights, and I'm sure something can
be worked out, or, decide that that particular piece is no longer needed
in VMS, take it out, and send said right holders on their way. If it
was important, replace it with new code.

>>>> If approval was given
>>>> all someone would really need to do is slap a new license header and
>>>> copyright statement at the top of each file HP hold copyrights for
>>>> and dump it on an FTP server somewhere.
>>> And even that would cost quite a bit and again, what would HP be getting
>>> out of it to justify the expense? Good will? Hardly. Most of the IT
>>> world is unaware of VMS and probably couldn't care less. That leaves
>>> a handful of C.O.V denizens who want to continue running it in their
>>> basement. hardly a justification for the cost or the potential liability.
>>>
>>>> It may just take some motivated people with contacts in HP to make
>>>> something happen. Considering the alternative I'd think it would
>>>> certainly be worth a try...
>>> Worth it to who? You? HP?

To me. That's enough for me.

>>>> Of course if everyone within the community is content to see OpenVMS
>>>> ultimately go extinct then without a doubt that is what will eventually
>>>> happen.
>>> One need only look at the fate of other similar products. Once again I
>>> have to bring up the PDP-11 OSes. Probably less than a tenth the number
>>> of lines of source code and all attempts to pry it from the hands of its
>>> owner since its demise have been fruitless. And guess who the top of
>>> the IP foodchain for these OSes is. :-)
>>>
>>> bill
>>>
>> I think that JF has been outdone. We have a new #1 ney-sayer ....
>>
>
> What's the difference between a nay-sayer and a realist? I spent many
> years fighting to keep VMS in the university environment. I went head-
> to-head with their management trying to explain why their "education"
> program was not going to fly. We all know where that ended. (I just
> asked at the datacenter today and all VMS is now gone from here, not
> just from the academic side.) I have been involved in a number of
> other cases where a desire exosted to save some piece of defunct
> software. None of them were ever succesful. In at least two cases
> I was directly involved in the current holder of the IP openly admited
> that they deliberately destroyed all copies of the sources.

A university environment might not be the best place to observe the
benefits of VMS. They don't care. They don't have critical processes
controlled by computers. You probably know this better than most.

[JKB]

Le Wed, 12 Jun 2013 17:13:50 +0000 (UTC),
Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> écrivait :

> On 2013-06-12, JKB <j...@koenigsberg.invalid> wrote:
>> Le Wed, 12 Jun 2013 06:12:58 -0700 (PDT),
>> Bob Gezelter <geze...@rlgsc.com> écrivait :
>>>
>>> JKB,
>>>
>>> There is vanishingly little of the OpenVMS source base that is architecture-specific to IA64. The vastly overwhelming majority of the "assembler" level code is written in MACRO-32, which is compiled for both Alpha and IA64.
>>
>> I know. I have the listing of 8.2.
>>
>>> Having been down both paths (re-writing vs modifying compilers) in the past, I can assure you that it is almost without a doubt more effective to do a port by redoing the translator than to do a recode.
>>>
>>> In this case, a port to x64 would require a new backend for the MACRO32 compiler, the BLISS compiler, and one or two others. Work, but not as challenging as a complete re-write.
>>>
>>> - Bob Gezelter, http://www.rlgsc.com
>>
>> Sure. But you suppose that HP will opensource OpenVMS. If you write
>> a clone, you can start to write this clone today. For information,
>> there is a Bliss compiler written as gcc front end.
>>
>
> There's also the fact that VMS requires much more hardware support than
> other operating systems do. Writing a new backend isn't going to do anything
> about that. :-)
>
> However (IIRC) you have said in the past that you have been able to use
> the microkernel design approach to emulate the multiple protection rings
> within your FreeVMS microkernel design.

Yes, I have.

[Simon Clubley]

On 2013-06-12, David Froble <da...@tsoft-inc.com> wrote:
>
> I'd think that HP has the right to sell, or give away, anything in VMS.

You would think wrong. Sorry. :-)

The contract might say that HP can freely use the source internally, but
only binaries can be shipped to customers. One perfectly valid reason for
this is that the source may contain implementation secrets that can only
be looked at under NDA.

There are plenty of examples in the open source world of similar issues
although in this case they tend to refer to interface specifications
instead of source code. However, the NDAs can still apply to really
basic things; for example you cannot get a interface specification out
of FTDI for it's USB serial port devices unless you sign a NDA.

Believe me, I've tried. :-(

(And the same issues also apply to a _full_ interface specification for
PL-2303 devices, for example, so it's not just FTDI locking down things
at this basic level of hardware.)

[Simon Clubley]

On 2013-06-12, Bill Gunshannon <bi...@server2.cs.uofs.edu> wrote:
>
> If it isn't a port of VMS why wold anyone here care? I thought VMS
> was what everyone wanted? TPU? There are hundreds of editors out
> there. TPU is just another. DLM? There has been a freely available
> DLM for at least a decade. Pieces of VMS are of little if any value.
> It is VMS as a system that has value and that everyone here wanted to
> see (or so I thought!)
>

I wouldn't mind seeing a full VMS style print and batch queue manager
implemented in Linux, especially if the printing side of things integrated
with CUPS in place of where DCPS is on VMS and if it implemented decent
support for forms management.

However, I also suspect that's the kind of thing which is more likely to
be written from scratch instead of been ported from the existing code base.

(And yes, I know there are commercial products, which cost a equally
commercial amount of money :-) available and offer much greater
functionality, but it would be nice to have something comparable to
VMS functionality freely available on Linux.)

[John Reagan]

"David Froble" wrote in message news:kpag12$iqc$1...@dont-email.me...

> If there are royalties being paid, HP surely knows to whom and why.

> I'd think that HP has the right to sell, or give away, anything in VMS.

I'll give you a real example. The DXML library (Digital eXtended Math
Library). Want to know why it isn't on Itanium? Some of the code used in
DXML is licensed (a one time fee, not any ongoing royalty situation) for
Alpha only. We couldn't legally ship it to customers on Itanium. While I
wasn't directly involved (I was one cubicle over), there was talk about
relicensing etc but something wasn't going to work out. I think the last
situation was that we might provide it in some partially-baked form and let
each customer go acquire the license/code to complete the library. I think
end-users could get a free license but vendors like HP had to do some other
form of license. Most of the time, I just turned up my headphones...

John

[John Wallace]

On Jun 12, 8:15 pm, Simon Clubley <clubley@remove_me.eisner.decus.org-

OK, the principle is understood, but are you (or is anybody) aware of
specific relevant examples where it applies (or even might apply) to
real VMS source code?

There are people round here with VMS listings. I'm not one of them,
but surely if it's been made obvious in the listings that there are
restrictions on the IP, someone might know by now. If something's been
deliberately left off the listings, someone might know by now.

I also quite like David Froble's earlier point along the lines of "if
royalties are payable, HP will know". On similar lines I suspect it's
also the case that if NDAs are applicable, HP's lawyers will know -
don't NDAs typically need renegotiation if a company changes hands, eg
DEC->CPQ->HP implies two sets of renegotiation?

And even if the NDA situation does apply, what is the practical
impact? FTDI need an NDA, that's understood. Does that stop people
developing Linux or Android or whatever drivers for FTDI? It might, if
they want to stay whiter than white. Surely much of the rest of the
world will carry on using whatever works, just as they do with drivers
and firmware supplied in binary-only formats?

Or have I missed something (it happens)?

[Simon Clubley]

On 2013-06-12, Stephen Hoffman <seao...@hoffmanlabs.invalid> wrote:
>
> Scroll down to the bottom of
><http://en.wikipedia.org/wiki/BLISS_(programming_language)> for what
> looks to be the download link.
>

I've now got it, thanks.

[Phillip Helbig---undress to reply]

In article <slrnkrgvq...@rayleigh.systella.fr>, JKB

<j...@koenigsberg.invalid> writes:

> Sure. But you suppose that HP will opensource OpenVMS. If you write
> a clone, you can start to write this clone today. For information,
> there is a Bliss compiler written as gcc front end.

But isn't freeVMS supposed to be a VMS clone? How many years or decades
has it been in development? How far along is it?

[Phillip Helbig---undress to reply]

In article <51b8b4ec$0$57828$c3e8da3$eb76...@news.astraweb.com>, JF
Mezei <jfmezei...@vaxination.ca> writes:

> Imagine logical names, with the LMF$SYSCLUSTER concepts allowing logical
> names to exists across Linux nodes.

People who appreciate that are supposed to be happy with Linux?

> And if HP were even to do the ports of those to Linux, it might help the
> remaining VMS custoemr base off VMS faster and start buying x86 servers
> from HP along with Linux support.

There seem to be enough moving off VMS already.

[Bill Gunshannon]

In article <kpag12$iqc$1...@dont-email.me>,

David Froble <da...@tsoft-inc.com> writes:
> Bill Gunshannon wrote:
>> In article <kpaa7u$f7s$1...@dont-email.me>,
>> David Froble <da...@tsoft-inc.com> writes:
>>> Bill Gunshannon wrote:
>>>> In article <d520f739-10eb-4dec...@googlegroups.com>,
>>>> dgso...@gmail.com writes:
>>>>> On Wednesday, June 12, 2013 2:44:37 PM UTC+12, Bill Gunshannon wrote:
>>>>>> In article <5gkk8a-...@nntp.local.net>,
>>>>>>
>>>>>> Single Stage to Orbit <alex....@munted.eu> writes:
>>>>>>
>>>>>>> With the news that HP is to take VMS out the back and put a bullet
>>>>>>> through its brains, can we hope that HP will open source the whole thing
>>>>>>> so that people can keep using it?=20
>>>>>>
>>>>>> Not a snowball's chance in hell.
>>>>> Is there any particular legal or technical reason why HP couldn't
>>>>> open-source it?
>>>> It would take an army of engineers and lawyers to go thru every single
>>>> line of code to determine that they actually have the right to do it.
>>>> Someone has to pay for that. HP is very unlikely to be willing to
>>>> accept that cost as they would be getting nothing in return and it
>>>> is just as easy to just destroy everything and let it die.
>
> I question this line of argument.
>
> If there are royalties being paid, HP surely knows to whom and why.

Royalties do not have to be paid for someone else to actually own
the IP. Just because HP had permission to use something, royalty
fee, does not mean they have the rights to pass that benefit on to
third parties.

>
> I'd think that HP has the right to sell, or give away, anything in VMS.

Not necessarily the case. Look at the infamous AT&T lawsuit. You would
have thought that BSD had the rights to give away the software they had
been developing. AT&T's lawyers thought otherwise. And they won.

> Now, the rights of whoever gets the software might be another issue.
> But it would be their issue, not HPs.

Not necessarily so. HP could be held liable for not protecting IP
that had been entrusted to them with the understanding that they
would, in fact, keep it safe.

> So in such a case there would not
> be this assumed cost to HP.

But can they afford to take the risk? What is it we keep hearing about
"due dilligence"? And, the easiest solution is to not release any of it.
That costs them nothing. Releasing it could cost a lot and contrary to
what people here seem to be saying (could people here actually be biased?)
There is no real gain for HP if they give VMS away.

>
>>>>> Assuming they haven't sold it they presumably still
>>>>> hold copyrights for most of the core platform.
>>>> An assumption that would take that army of engineers and lawyers to
>>>> determine.
>
> Why? If someone has a problem with the users of the open sourced VMS,
> they can inform said users of their rights, and I'm sure something can
> be worked out, or, decide that that particular piece is no longer needed
> in VMS, take it out, and send said right holders on their way. If it
> was important, replace it with new code.

Or just sue everyone for infringement of their IP. That's how lawyers
make their money. And HP has a lot of high-priced lawyers on staff to
make sure that doesn't happen. And the easiest way to do that is to
just say "no". Let's go back to the HP Educational Use License issue.
After much discussion with HP they sent me a copy of what they were
proposing. I took it to the University Counsle. Her immediate response
was, "We can't sign that." That is the crux of the matter. The easiest
path for HP's lawyers to take is to advise burying all of it and never
letting it see the light of day again. I am waiting for someone to come
up with any possible real benefit to HP from risking possible liability
by releasing it into the wild.

>
>>>>> If approval was given
>>>>> all someone would really need to do is slap a new license header and
>>>>> copyright statement at the top of each file HP hold copyrights for
>>>>> and dump it on an FTP server somewhere.
>>>> And even that would cost quite a bit and again, what would HP be getting
>>>> out of it to justify the expense? Good will? Hardly. Most of the IT
>>>> world is unaware of VMS and probably couldn't care less. That leaves
>>>> a handful of C.O.V denizens who want to continue running it in their
>>>> basement. hardly a justification for the cost or the potential liability.
>>>>
>>>>> It may just take some motivated people with contacts in HP to make
>>>>> something happen. Considering the alternative I'd think it would
>>>>> certainly be worth a try...
>>>> Worth it to who? You? HP?
>
> To me. That's enough for me.

Yes, but neither own it nor have the potential liability from releasing
it. Seriously, what is in it for HP? Good will? With who?

>
>>>>> Of course if everyone within the community is content to see OpenVMS
>>>>> ultimately go extinct then without a doubt that is what will eventually
>>>>> happen.
>>>> One need only look at the fate of other similar products. Once again I
>>>> have to bring up the PDP-11 OSes. Probably less than a tenth the number
>>>> of lines of source code and all attempts to pry it from the hands of its
>>>> owner since its demise have been fruitless. And guess who the top of
>>>> the IP foodchain for these OSes is. :-)
>>>>
>>>> bill
>>>>
>>> I think that JF has been outdone. We have a new #1 ney-sayer ....
>>>
>>
>> What's the difference between a nay-sayer and a realist? I spent many
>> years fighting to keep VMS in the university environment. I went head-
>> to-head with their management trying to explain why their "education"
>> program was not going to fly. We all know where that ended. (I just
>> asked at the datacenter today and all VMS is now gone from here, not
>> just from the academic side.) I have been involved in a number of
>> other cases where a desire exosted to save some piece of defunct
>> software. None of them were ever succesful. In at least two cases
>> I was directly involved in the current holder of the IP openly admited
>> that they deliberately destroyed all copies of the sources.
>
> A university environment might not be the best place to observe the
> benefits of VMS. They don't care. They don't have critical processes
> controlled by computers. You probably know this better than most.

Well, many years ago people here were constantly talking about the
potential good of having VMS still present in academia. But that
totally aside, do you realize how many universities ran their entire
enterprise on VMS? Banner was one of the most common administrative
systems and it ran on VMS. Oracle? Long after academic use had all
but disappeared here we still had a room full of VMS boxes without
which the university cold not function. Academic institutions have
critical systems, too. And even non-profit universities are multi-
million dollar operations.

bill

>
>> Go ahead, get your hopes up. Sometimes it takes getting knocked down
>> real hard to get a point across.
>>
>>> VMS is more than just a bunch of hobbyists. But you know that. You're
>>> just having fun. why don't you go change the oil in the MG ??
>>
>> Is it? At this point, anyone who is not just a hobbyist who is betting
>> their business on VMS is a fool. It's not like the writting hasn't been
>> on the wall for a long time.
>>
>> As for the MG, fuel pump is on the way and I expect to be driving it
>> in about two weeks. I would expect everyone here to be an MG owner.
>> We all like antique systems and we are all gluttons for punishment.
>>
>> bill
>>
>>
>>

[Simon Clubley]

On 2013-06-12, John Wallace <johnwa...@gmail.com> wrote:
>
> OK, the principle is understood, but are you (or is anybody) aware of
> specific relevant examples where it applies (or even might apply) to
> real VMS source code?
>

Apart from the example John has just posted, the other obvious candidates
are the drivers for graphics hardware, given what happens in the open
source world.

You can also extend that to hardware drivers in general if DEC/CPQ/HP
bought in the chipsets instead of manufacturing them in house.

> There are people round here with VMS listings. I'm not one of them,
> but surely if it's been made obvious in the listings that there are
> restrictions on the IP, someone might know by now. If something's been
> deliberately left off the listings, someone might know by now.
>
> I also quite like David Froble's earlier point along the lines of "if
> royalties are payable, HP will know". On similar lines I suspect it's
> also the case that if NDAs are applicable, HP's lawyers will know -
> don't NDAs typically need renegotiation if a company changes hands, eg
> DEC->CPQ->HP implies two sets of renegotiation?
>
> And even if the NDA situation does apply, what is the practical
> impact? FTDI need an NDA, that's understood. Does that stop people
> developing Linux or Android or whatever drivers for FTDI? It might, if
> they want to stay whiter than white. Surely much of the rest of the
> world will carry on using whatever works, just as they do with drivers
> and firmware supplied in binary-only formats?
>
> Or have I missed something (it happens)?

The practical impact is that all the code needs to be audited to make
sure it doesn't fall under a NDA before HP can be sure it's safe to
release. That takes time and money. The code can be re-written after
it's been identified, but it needs to be identified first.

I'm not trying to be a "naysayer" here (I would love to see VMS open
sourced, just like I would the PDP-11 operating systems), but I'm just
trying to point out the possible issues in the VMS codebase after
encountering similar issues in the open source world.

[Stephen Hoffman]

On 2013-06-12 19:43:34 +0000, John Wallace said:

> OK, the principle is understood, but are you (or is anybody) aware of
> specific relevant examples where it applies (or even might apply) to
> real VMS source code?
>
> There are people round here with VMS listings. I'm not one of them,
> but surely if it's been made obvious in the listings that there are

> restrictions on the IP, someone might know by now....

>
> Or have I missed something (it happens)?

The source listings are expurgated. Not all of the code of VMS is
included in those kits.

While the bulk of the source code is copyright by HP (or DEC or
Compaq), there are hunks of the base-VMS kit that aren't HP-owned code.
The IntraServer SCSI driver comes to mind here, as that component
"advertises" itself at VMS bootstrap. Common Data Security
Architecture (CDSA) is another hunk of third-party code involved here,
too, and that stuff is discussed in the VMS documentation, and you can
easily find CDSA source code with the "Copyright (c) 1995-2000 Intel
Corporation. All rights reserved." statement in its include files.
There's other third-party stuff around, too.

I am also here referring to the base VMS kit. Not to X. Not to IP.
Not to the compilers and tools and procedures necessary to build VMS.
Not to any other layered products. Etc.

Any hypothetical open-sourcing of VMS would necessarily involve various
managerial decisions, a code review or three, and lawyers.

[John Wallace]

On Jun 12, 8:40 pm, "John Reagan" <johnrrea...@earthlink.net> wrote:
> "David Froble"  wrote in messagenews:kpag12$iqc$1...@dont-email.me...

I know that one from ages ago. Most folk won't, so...

DXML (and its renamed successor, CXML) started life as a separately
orderable package providing an Alpha-optimised math library, with
Alpha-friendly (often specifically cache-friendly) versions of various
math routines in relatively common use, intended to make the most of
Alpha's performance with floating point arithmetic. The basic details
are (obviously) in the SPDs.

The SPD for HP Fortran says that CXML was eventually bundled with HP
Fortran for VMS, e.g.
http://h71000.www7.hp.com/doc/spdfortran.pdf though I don't remember
anything that would make it Fortran specific (this may just have been
a mechanism to simplify licencing and ordering).

There was also the similarly named but largely unrelated DPML (aka
DIGITAL Portable Math Library, a different set of lower level
numbercrunching routines: "DPML was developed to provide a common set
of routines that support many of the common mathematical functions
across a wide variety of operating systems, hardware architectures,
and languages.", says the Fine Manual).

DXML might indeed have been a real example once, but is it still a
relevant example?

Is any of this relevant to real paying customers now?

[JKB]

Le Wed, 12 Jun 2013 19:49:15 +0000 (UTC),
Phillip Helbig---undress to reply <hel...@astro.multiCLOTHESvax.de> écrivait :

Sorry, but I cannot understand this argument. On mailing list, I
have _today_ more than 1000 real subscribers. But all subscribers
are in read only mode.

Until 0.3, we are only _two_ active developers. If there were more
than two overbooked developers FreeVMS should be usable _today_. But
I think that all potential users wait for a completed FreeVMS and
don't want to contribute.

With new L4 architecture, all drivers can be independently written.
You don't have to understand all internal kernel capabilities to
write a driver. If someone want to write some pieces of code, he will
be welcome. Today, FreeVMS waits for a pager (pager is started but
does not anything).

To be usable, we need a running kernel with a subset of OpenVMS
features. Not all in a first time. But if anyone contributes, this
project cannot stay alive.

[David Froble]

Stephen Hoffman wrote:
> On 2013-06-12 19:43:34 +0000, John Wallace said:
>
>> OK, the principle is understood, but are you (or is anybody) aware of
>> specific relevant examples where it applies (or even might apply) to
>> real VMS source code?
>>
>> There are people round here with VMS listings. I'm not one of them,
>> but surely if it's been made obvious in the listings that there are
>> restrictions on the IP, someone might know by now....
>>
>> Or have I missed something (it happens)?
>
> The source listings are expurgated. Not all of the code of VMS is
> included in those kits.

Ok, here's a question. Could it be assumed that the source listings are
free of any IP and anything else that HP could not "open source" or
otherwise make available? If so, then question #2 becomes, how much of
VMS is there? Would it be a good start on a non-HP VMS?

The last set of sources I had was V4 on microfiche. I have no idea what
might be on later releases on CD. Nor am I aware of the circumstances
under which the sources were released. Were people told anything about
actually trying to build the sources into an OS? I remember some being
given for free at past DECUS events.

Regardless, supporting VMS into the future is a non-starter without the
compilers. For example, a port to x86 would need the Macro-32 compiler
and Bliss compiler, just as a starting point. Actually, for any
re-build of the OS, even on IA-64 or Alpha or VAX would need the compilers.

> While the bulk of the source code is copyright by HP (or DEC or Compaq),
> there are hunks of the base-VMS kit that aren't HP-owned code. The
> IntraServer SCSI driver comes to mind here, as that component
> "advertises" itself at VMS bootstrap. Common Data Security Architecture
> (CDSA) is another hunk of third-party code involved here, too, and that
> stuff is discussed in the VMS documentation, and you can easily find
> CDSA source code with the "Copyright (c) 1995-2000 Intel Corporation.
> All rights reserved." statement in its include files. There's other
> third-party stuff around, too.
>
> I am also here referring to the base VMS kit. Not to X. Not to IP.
> Not to the compilers and tools and procedures necessary to build VMS.
> Not to any other layered products. Etc.
>
> Any hypothetical open-sourcing of VMS would necessarily involve various
> managerial decisions, a code review or three, and lawyers.
>
>

While it may make any attempt more difficult, getting any part of VMS
that doesn't have strings attached would be much better than starting
from scratch. I'd guess that starting with nothing would never be
attempted. But with some starting point, perhaps there might be enough
interest to get something started.

The way I see it, if there are parts we could not acquire, then there
would have to be a decision as to whether those parts would be required,
at first, or ever.

[Scott Dorsey]

JKB <j...@koenigsberg.invalid> wrote:
>
> Sure. But I think that today, if you want run OpenVMS on new
> hardware, you have to buy an IA64. You can try to buy nuVAX or an
> AXP emulator, but it's not a solution.

I hate to say it, but if I had to maintain a system for the long term, I
think getting parts and support for a 25-year-old Microvax would be easier
than getting support for a 5-year-old iTanium.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Craig A. Berry]

In article <slrnkrgvq...@rayleigh.systella.fr>,

JKB <j...@koenigsberg.invalid> wrote:

> Sure. But you suppose that HP will opensource OpenVMS. If you write
> a clone, you can start to write this clone today. For information,
> there is a Bliss compiler written as gcc front end.

That's a non sequitur. If you are writing your own from scratch, there
is no reason to use BLISS.

[John Reagan]

"John Wallace" wrote in message
news:8a825883-b2b1-4e8f...@c7g2000vbv.googlegroups.com...

> DXML might indeed have been a real example once, but is it still a
> relevant example?

> Is any of this relevant to real paying customers now?

I was making a point that there is code that:

- isn't under royalty so there is nobody keeping track of it
- Digital/Compaq/HP uses it with a license for a specific system
- I doubt that anybody left besides me knows the details without looking at
the code to check the copyright statements and I'm not sure if the copyright
statements in the source are enough to describe the situation

Cases such as this (this might be the only one but it might not be the only
one) require manual inspection and research if you wish to discuss open
sourcing the code. You might simply say: "if you see any non
Digital/Compaq/HP copyright, then you simply cannot open source it without
further research"

[Michael Kraemer]

Bill Gunshannon schrieb:

> Not necessarily the case. Look at the infamous AT&T lawsuit. You would
> have thought that BSD had the rights to give away the software they had
> been developing. AT&T's lawyers thought otherwise. And they won.

Another case in point:
SCO's feud with IBM over Unix IP allegedly going
Linux via AIX. At some point SCO wanted to inspect
all of AIX' sources. And they got it.
So for being safe from IP or patent trolls,
better be sure to release 100% "clean" sources.

[Arne Vajhøj]

On 6/12/2013 2:57 PM, David Froble wrote:
> Bill Gunshannon wrote:
>> In article <kpaa7u$f7s$1...@dont-email.me>,
>> David Froble <da...@tsoft-inc.com> writes:
>>> Bill Gunshannon wrote:
>>>> In article <d520f739-10eb-4dec...@googlegroups.com>,
>>>> dgso...@gmail.com writes:
>>>>> On Wednesday, June 12, 2013 2:44:37 PM UTC+12, Bill Gunshannon wrote:
>>>>>> In article <5gkk8a-...@nntp.local.net>,
>>>>>>
>>>>>> Single Stage to Orbit <alex....@munted.eu> writes:
>>>>>>
>>>>>>> With the news that HP is to take VMS out the back and put a bullet
>>>>>>> through its brains, can we hope that HP will open source the
>>>>>>> whole thing
>>>>>>> so that people can keep using it?=20
>>>>>>
>>>>>> Not a snowball's chance in hell.
>>>>> Is there any particular legal or technical reason why HP couldn't
>>>>> open-source it?
>>>> It would take an army of engineers and lawyers to go thru every single
>>>> line of code to determine that they actually have the right to do it.
>>>> Someone has to pay for that. HP is very unlikely to be willing to
>>>> accept that cost as they would be getting nothing in return and it
>>>> is just as easy to just destroy everything and let it die.
>
> I question this line of argument.
>
> If there are royalties being paid, HP surely knows to whom and why.

But if not ....

> I'd think that HP has the right to sell, or give away, anything in VMS.

Why? If they are using anything from third party it is very unlikely
that they are licensed to release it as open source.

> Now, the rights of whoever gets the software might be another issue.
> But it would be their issue, not HPs. So in such a case there would not
> be this assumed cost to HP.
>
>>>>> Assuming they haven't sold it they presumably still
>>>>> hold copyrights for most of the core platform.
>>>> An assumption that would take that army of engineers and lawyers to
>>>> determine.
>
> Why? If someone has a problem with the users of the open sourced VMS,
> they can inform said users of their rights, and I'm sure something can
> be worked out, or, decide that that particular piece is no longer needed
> in VMS, take it out, and send said right holders on their way. If it
> was important, replace it with new code.

Or they could sue HP because HP got money.

Why would HP take the risk?

Arne

[Arne Vajhøj]

On 6/12/2013 3:43 PM, John Wallace wrote:
> On Jun 12, 8:15 pm, Simon Clubley <clubley@remove_me.eisner.decus.org-
> Earth.UFP> wrote:
>> On 2013-06-12, David Froble <da...@tsoft-inc.com> wrote:
>>
>>
>>
>>> I'd think that HP has the right to sell, or give away, anything in VMS.
>>
>> You would think wrong. Sorry. :-)
>>
>> The contract might say that HP can freely use the source internally, but
>> only binaries can be shipped to customers. One perfectly valid reason for
>> this is that the source may contain implementation secrets that can only
>> be looked at under NDA.
>>
>> There are plenty of examples in the open source world of similar issues
>> although in this case they tend to refer to interface specifications
>> instead of source code. However, the NDAs can still apply to really
>> basic things; for example you cannot get a interface specification out
>> of FTDI for it's USB serial port devices unless you sign a NDA.
>>
>> Believe me, I've tried. :-(
>>
>> (And the same issues also apply to a _full_ interface specification for
>> PL-2303 devices, for example, so it's not just FTDI locking down things
>> at this basic level of hardware.)
>

> OK, the principle is understood, but are you (or is anybody) aware of
> specific relevant examples where it applies (or even might apply) to
> real VMS source code?

Not me.

But the chance that there are third party source code in a code base
as big of VMS is pretty big.

> There are people round here with VMS listings. I'm not one of them,
> but surely if it's been made obvious in the listings that there are
> restrictions on the IP, someone might know by now. If something's been
> deliberately left off the listings, someone might know by now.

Note that certain pieces may have been left out of the traditional
listings dist.

I seem to remember that there existed a real buildable source dist. If
I remember correct then that should have everything.

Arne

[Arne Vajhøj]

On 6/12/2013 1:50 PM, John Wallace wrote:
> On Jun 11, 11:21 pm, Single Stage to Orbit <alex.bu...@munted.eu>

> wrote:
>> With the news that HP is to take VMS out the back and put a bullet
>> through its brains, can we hope that HP will open source the whole thing
>> so that people can keep using it?
>

> I'm told it's now the 21st century. Allegedly Web 2.0 and the like
> provide mechanisms to find out how interesting a project like this
> would be, and whether there's any real potential backing out there.
>
> Who knows enough about crowdsourcing (kickstarter etc) to go about
> setting something up?
>
> Who thinks enough people would be willing to commit (donate?) a
> worthwhile sum of money to get something going far enough to get HP's
> attention?
>
> I know nothing about this stuff, I don't even know whether I think
> this might be a good idea let alone whether I think it stands any
> chance of success, but in theory it ought to be a possibility in the
> age of LinkedIn communities and crowdsourcing and other such
> trendiness.
>
> If nothing else, it might help answer the "worth it? to who?"
> question. It still leaves a few challenges left to overcome (e.g. the
> compiler backends Bob mentions, and maybe the copyright issues
> routinely mentioned, though for those there still aren't any
> functional showstoppers I'm aware of).

Based on the fact that for the last 10 years:

number of persons complaining over lack of software on VMS has
been an order of magnitude higher than the number of people actually
developing or porting software to VMS

then I would be very skeptical.

Arne

[Arne Vajhøj]

On 6/11/2013 9:40 PM, dgso...@gmail.com wrote:
> On Wednesday, June 12, 2013 12:28:02 PM UTC+12, Arne Vajhøj wrote:

>> On 6/11/2013 6:21 PM, Single Stage to Orbit wrote:
>>> With the news that HP is to take VMS out the back and put a
>>> bullet
>>> through its brains, can we hope that HP will open source the
>>> whole thing
>>> so that people can keep using it?
>>

>> You can hope.
>>
>> But it does not seem that likely.
>
> They don't have to open-source the *whole* thing. Just open-sourcing
> the bits they hold the copyrights on (which I expect is the majority
> of the core system) may be enough. Replacements can probably be
> written for whatever HP can't release (as happened with OpenJDK,
> Illumos, Haiku, etc).

OpenJDK is a good example.

SUN spent significant resources creating binary modules to fill out
the gaps in source.

And Redhat spent significant resources creating open source
replacements.

Arne

[Arne Vajhøj]

On 6/11/2013 9:53 PM, JF Mezei wrote:

> On 13-06-11 20:28, Arne Vajhøj wrote:
>> You can hope.
>> But it does not seem that likely.
>

> HP bought Palm and a short time later abandonned its investment and
> announced it would open source the abandonned OS.
>
> This was very proprietary OS and I suspect encumbered with a lot of
> stuff. And HP never made any money from it, so justifying the spending
> to determine what could and couldn,T be made open source was even less
> justified than similar process for VMS which made HP lots of money over
> the years.

Historic revenue has no relevance on decisions for the future.

They did open source WebOS, but there are a few relevant points
here:
* they just bought WebOS 10 months before open sourcing, so they
could reuse their acquisition due diligence results for the open
sourcing
* at the time of open sourcing it was only 2 years old so a lot easier
to find copyright holders
* It does not have its own kernel - it uses Linux kernel - so that
part was already open source
* HP may actually make money from the open sourcing (assuming that
LG actually pays for licensing from HP)

VMS is 36 years old. It is not build on an open source kernel. And
I seriously doubt that a south korean giant will be interested in
buying a license.

Arne

[Arne Vajhøj]

On 6/12/2013 3:52 PM, Bill Gunshannon wrote:
> In article <kpag12$iqc$1...@dont-email.me>,
> David Froble <da...@tsoft-inc.com> writes:
>> I'd think that HP has the right to sell, or give away, anything in VMS.
>
> Not necessarily the case. Look at the infamous AT&T lawsuit. You would
> have thought that BSD had the rights to give away the software they had
> been developing. AT&T's lawyers thought otherwise. And they won.

Are you talking about USL vs BSDi that was settled out of court?

>> Now, the rights of whoever gets the software might be another issue.
>> But it would be their issue, not HPs.
>
> Not necessarily so. HP could be held liable for not protecting IP
> that had been entrusted to them with the understanding that they
> would, in fact, keep it safe.

Yes.

>> So in such a case there would not
>> be this assumed cost to HP.
>
> But can they afford to take the risk? What is it we keep hearing about
> "due dilligence"? And, the easiest solution is to not release any of it.
> That costs them nothing. Releasing it could cost a lot and contrary to
> what people here seem to be saying (could people here actually be biased?)
> There is no real gain for HP if they give VMS away.

Yes.

Arne

[Arne Vajhøj]

On 6/12/2013 1:50 PM, JF Mezei wrote:

> On 13-06-12 10:26, Bill Gunshannon wrote:
>> Someone has to pay for that. HP is very unlikely to be willing to
>> accept that cost as they would be getting nothing in return and it
>> is just as easy to just destroy everything and let it die.
>

> Then why did HP decide to go through that trouble for its PalmOS/WebOS ?

They had most likely gone through everything when they acquired it 10
months earlier.

And it turned out that LG was willing to license it after it was
open sourced. I assume that mean that they paid money.

Low cost because cost already paid for and extra revenue.

Arne

[Arne Vajhøj]

On 6/12/2013 9:12 AM, Bob Gezelter wrote:
> Having been down both paths (re-writing vs modifying compilers) in
> the past, I can assure you that it is almost without a doubt more
> effective to do a port by redoing the translator than to do a
> recode.

For initial cost absolutely.

For long term cost maybe.

Maintenance cost for Macro-32 will be a lot higher than for
something a bit more high level.

Arne

[Richard B. Gilbert]

On 6/11/2013 6:21 PM, Single Stage to Orbit wrote:
> With the news that HP is to take VMS out the back and put a bullet
> through its brains, can we hope that HP will open source the whole thing
> so that people can keep using it?
>

I don't believe that they CAN do that. The problem is that H-P does NOT
own all the intellectual property involved.

[dgso...@gmail.com]

The don't have to release ALL of it. Releasing just the bits they DO own would be far far better than releasing nothing at all.

Java - only the bits owned by Sun were released. The rest was replaced by the community (see: OpenJDK)

Solaris - Again, not all of it was released. What wasn't was replaced by the community (see: Illumos)

BeOS - only Tracker, the desktop environment, was released IIRC (compiler and text-mode software was all GNU stuff). The rest of the operating system (kernel, drivers, application frameworks, windowing system, graphical utilities, etc) was rewritten by the community (see: Haiku Project)

Tru64 - only AdvFS was released. AFAIK no one has attempted to rewrite the rest of the OS yet :)

Obviously the more HP can release the easier it will be to keep OpenVMS alive. But if they release nothing then OpenVMS probably has no future.

So saving OpenVMS probably is possible if HP can be convinced to open source a useful amount of it. Otherwise whatever they do release may at least help people port their software to other platforms.

Instead of everyone discussing how hard it is, listing all the reasons why HP might not do it, etc. How about we seriously discuss how we could actually make something happen. What are the options? Has anyone actually tried contacting people at HP? Talking to people who make decisions?

Remember that HP still has a bunch of people on staff looking after OpenVMS. People who might be able to help. Any attempt to save it is best made before all these people are reassigned.

[JF Mezei]

On 13-06-12 21:01, Arne Vajhøj wrote:

> But the chance that there are third party source code in a code base
> as big of VMS is pretty big.

The DHCP server used to be by a company called "Join". It no longer
exists. So who would sue HP over such rights ? And what if HP had full
rights, including open sourcing upon Join ceasing to exist ?


If creditors inrehited the IP from Join, wouldn't it be their
responsability to knock on HP's door and introduce themselves with:
"hey, from now on, you answer to us whenever you need to do something
with the Join DHCP server". ??


Where I could see demand for open source would be the USA military. If
those planes with VMS nodes in them are still flying with that
equipment, the military may require source in exchange for HP ending VMS
development.

That assumes those planes still use VMS on Alphas
.

[JF Mezei]

On 13-06-12 23:16, dgso...@gmail.com wrote:

> Remember that HP still has a bunch of people on staff looking after OpenVMS. People who might be able to help. Any attempt to save it is best made before all these people are reassigned.
>

Perhaps Mr Keith Paris could go to his bosses and ask how big a
community voice would be required to get HP to consider open sourcing
VMS. This may give the community some sort of idea whether such effort
would be realistic or not.

The problem here is that HP has not announced a formal EOL for VMS, so
they are not about to start openly talking about open sourcing a product
that isn't officially dead yet.

So all this would have to be informal and under the table util (and IF)
HP ever decides it has the guys to make the formal EOL announcement.

Remember that it isn't a given that HP will ever announce EOL for VMS.

[David Froble]

dgso...@gmail.com wrote:

> Instead of everyone discussing how hard it is, listing all the
> reasons why HP might not do it, etc. How about we seriously discuss
> how we could actually make something happen. What are the options?
> Has anyone actually tried contacting people at HP? Talking to people
> who make decisions?

Yeah, I was thinking about saying something like this. For those who
want to raise obstacles, go ahead, but some of us just might be able to
do something.

I don't know that anything CAN be done, but if somebody doesn't try, for
sure nothing will happen.

[David Froble]

If it doesn't jeprodize his job, perhaps Keith could sound out some
people at HP. But right about now, I have to wonder what there is to
jeprodize.

I see that you think that things have to be A or B, X or Y, but that
isn't so. HP could license a consortium to take whatever can be
released and develop their own product, while still selling and
supporting VMS themselves. Face it, if we get part of the whole, we'd
be lucky to have a ready product in the next 7 years. If the goal is
something else, such as a port to x86, that will not be trivial.

Maybe Brian should start sending John Reagan a monthly case of his
favorite brew. A primary need for any port will be the Macro-32
compiler, and the Bliss compiler ....

I'm starting to think about that documentation I got laying around on
Bliss ....

[JF Mezei]

On 13-06-13 02:08, David Froble wrote:

> Maybe Brian should start sending John Reagan a monthly case of his
> favorite brew. A primary need for any port will be the Macro-32
> compiler, and the Bliss compiler ....

Me thinks the x86 version of Macro64 was done many years ago.
Wasn't there a Bliss for x86 rumoured done as well ?

And me thinks Hoff is quietly laughing while reading all this
speculation from an x86 laptop running VMS natively in his living room.
Or perhaps he is crying because he is bound to secrecy and can't
announce that he's already ported VMS.

And we'll never know because HP isn't going to fess up to the fact that
the pilot project to evaluate porting VMS to x86 went quite far and
would be a no brainer to complete if the original engineering team were
reconstituted.

Looks guys, it's over. We can laugh about it. We have to laugh about it.
Not to rain on your parade but HP has planned this a long time ago, and
it isn't a small number of grumpy ex VMS customers (hobbyists) who will
change HP's plans.

[David Goodwin]

In article <51b97b77$0$17011$c3e8da3$76a7...@news.astraweb.com>,
jfmezei...@vaxination.ca says...

HP has no particular reason to scheme away carefully plotting the demise
of OpenVMS.

They're a company. Their purpose is to make money. What ever they do
should be seen in that light. If they canceled an x86 port it is because
they considered it the most profitable (directly or indirectly) thing to
do.

Remember who they'd be up against on the x86 platform. Long term even
Microsoft is going to have a hard time competing with Linux - OpenVMS
wouldn't really stand much of a chance. Especially as a proprietary
operating system with a limited selection of 3rd party software mostly
ported from what would be their main competition. Development and
support costs would be higher too due to the larger variety of hardware
(there is more than one reason why MacOS X is only licensed for Apple
hardware). If there ever was an x86 port they likely canceled it simply
because they felt it was not commercially viable. Same reason why they
probably won't bother porting HP-UX.

As for why OpenVMS was EOLd now? HP-UX seems like it would be in a worse
boat - it competes directly with Linux. What is there really to stop
people replacing their IA64 HP-UX boxen with x86 Linux? If IA64s fate
depends on HP-UX then the platform probably doesn't have much time left.
HP wouldn't have long to recover the costs of porting OpenVMS to Poulson
- especially if a large number of their customers haven't even migrated
from Alpha or VAX gear yet.

Again, I'd be surprised if there was a plot to kill OpenVMS. They just
think they have more of a future selling x86 Linux boxen. Nothing
personal - just business.

[Paul Sture]

In article <51b9264a$0$32114$1472...@news.sunsite.dk>,

Arne Vajhøj <ar...@vajhoej.dk> wrote:

> On 6/11/2013 9:53 PM, JF Mezei wrote:
> > On 13-06-11 20:28, Arne Vajhøj wrote:
> >> You can hope.
> >> But it does not seem that likely.
> >
> > HP bought Palm and a short time later abandonned its investment and
> > announced it would open source the abandonned OS.
> >
> > This was very proprietary OS and I suspect encumbered with a lot of
> > stuff. And HP never made any money from it, so justifying the spending
> > to determine what could and couldn,T be made open source was even less
> > justified than similar process for VMS which made HP lots of money over
> > the years.
>
> Historic revenue has no relevance on decisions for the future.
>
> They did open source WebOS, but there are a few relevant points
> here:
> * they just bought WebOS 10 months before open sourcing, so they
> could reuse their acquisition due diligence results for the open
> sourcing
> * at the time of open sourcing it was only 2 years old so a lot easier
> to find copyright holders
> * It does not have its own kernel - it uses Linux kernel - so that
> part was already open source
> * HP may actually make money from the open sourcing (assuming that
> LG actually pays for licensing from HP)

There was a lot of interest in WebOS too, from what I gather.

And the other part of WebOS being a recent acquisition was that it
allowed them to write a large chunk off their books and save a load of
tax. This one always appeals to the beancounters.

--
Paul Sture

IBM's Thomas J. Watson predicted a "world market for maybe five computers".
Given the way this whole Cloud thing is going, he might have been extremely
prescient.

[John Wallace]

On Jun 13, 11:22 am, David Goodwin <dgsof...@gmail.com> wrote:
> In article <51b97b77$0$17011$c3e8da3$76a7c...@news.astraweb.com>,
> jfmezei.spam...@vaxination.ca says...

Why?

The obvious sensible thing to do (ie probably the one least likely to
happen) would be target VMS development, qualification, and support at
a select few Proliant-class boxes. Do that and the hardware-related
design, qualification and support costs are paid, there's 'only' the
VMS-specific costs to pick up. If as a bonus VMS happens to run
(unsupported) on the smaller x86 boxes, e.g. for cheap+cheerful dev
+test purposes, that's a bonus.

But that might involve people in HQ (Intel and HP) admitting that IA64
might not have been a bright idea. It was a long time ago though.

"What is there really to stop people replacing their IA64 HP-UX boxen
with x86 Linux? "

The "who to sue" question seems to be considered important in some
places, rightly or wrongly.

"If IA64s fate depends on HP-UX then the platform probably doesn't
have much time left."

Living on borrowed time.

[Bill Gunshannon]

In article <51b95574$0$63709$c3e8da3$1e92...@news.astraweb.com>,

JF Mezei <jfmezei...@vaxination.ca> writes:
> On 13-06-12 21:01, Arne Vajhøj wrote:
>
>> But the chance that there are third party source code in a code base
>> as big of VMS is pretty big.
>
> The DHCP server used to be by a company called "Join". It no longer
> exists. So who would sue HP over such rights ?

Not knowing only increases the risk. That's why lawyers tend to be pretty
good at risk assessment. They always look at the dark side. (some say I
do, too, actually.)

> And what if HP had full
> rights, including open sourcing upon Join ceasing to exist ?

They would have those rights only if it was explicitly written in the
contract with Join. Have you ever seen a contract written like that?
Does any company put a contingency in their contracts that assumes their
demise? none that I have ever seen.

>
>
> If creditors inrehited the IP from Join, wouldn't it be their
> responsability to knock on HP's door and introduce themselves with:
> "hey, from now on, you answer to us whenever you need to do something
> with the Join DHCP server". ??

Not necessarily. They could happily wait around holding on to their
licenses, wait for someone (like HP) to violate the license and then
seek monetary damages. Sounds like a good way to milk that cow. And
it keeps lawyers in work.

>
>
> Where I could see demand for open source would be the USA military.

You might think that, but my (rather extensive) experience with DOD
IT demonstrated quite the opposite. While they seem aware of open
source, they still fail to see the advantage in using it.

> If
> those planes with VMS nodes in them are still flying with that
> equipment, the military may require source in exchange for HP ending VMS
> development.

Actually, the military has little to do with JStars beyond
operating it. Maintenance is totally done by the contractor,
who may have full access to the source as part of their
contract with HP. Can't say for sure as I haven't seen the
contract but I know we had similar agreements with vendors
like Prime back in my days as a DOD contractor. (I maintained
all the software on a bunch of Prime 50 Series Minis and I had
full source code for the OS and all layered products.)

>
> That assumes those planes still use VMS on Alphas.

I don't know where it stands but the last I saw about Jstars
there was talk of moving it all to Windows based systems. Take
up less space, consume less power, cost considerable less money.

bill

--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bill...@cs.scranton.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>

[Bill Gunshannon]

In article <51b9270e$0$32114$1472...@news.sunsite.dk>,

Arne Vajhøj <ar...@vajhoej.dk> writes:
> On 6/12/2013 3:52 PM, Bill Gunshannon wrote:
>> In article <kpag12$iqc$1...@dont-email.me>,
>> David Froble <da...@tsoft-inc.com> writes:
>>> I'd think that HP has the right to sell, or give away, anything in VMS.
>>
>> Not necessarily the case. Look at the infamous AT&T lawsuit. You would
>> have thought that BSD had the rights to give away the software they had
>> been developing. AT&T's lawyers thought otherwise. And they won.
>
> Are you talking about USL vs BSDi that was settled out of court?

Yes, it was. But that doesn't mean it didn't still cost BSDi a lot of
money in lawyers and court fees before an agreement was reached. (I
believe one of the reasons USL decided to settle out of court was the
threat of a counter suit from BSD over the removal of their copyright
notices from other code in the basic Unix distribution, which I also
mentioned earlier.)

>
>>> Now, the rights of whoever gets the software might be another issue.
>>> But it would be their issue, not HPs.
>>
>> Not necessarily so. HP could be held liable for not protecting IP
>> that had been entrusted to them with the understanding that they
>> would, in fact, keep it safe.
>
> Yes.
>
>>> So in such a case there would not
>>> be this assumed cost to HP.
>>
>> But can they afford to take the risk? What is it we keep hearing about
>> "due dilligence"? And, the easiest solution is to not release any of it.
>> That costs them nothing. Releasing it could cost a lot and contrary to
>> what people here seem to be saying (could people here actually be biased?)
>> There is no real gain for HP if they give VMS away.
>
> Yes.
>

[Bill Gunshannon]

In article <1e4252b4-ff19-49d5...@googlegroups.com>,

dgso...@gmail.com writes:
> On Thursday, June 13, 2013 2:41:36 PM UTC+12, Richard B. Gilbert wrote:
>> On 6/11/2013 6:21 PM, Single Stage to Orbit wrote:
>>
>> > With the news that HP is to take VMS out the back and put a bullet
>> > through its brains, can we hope that HP will open source the whole thing
>> > so that people can keep using it?
>>
>> I don't believe that they CAN do that. The problem is that H-P does NOT
>> own all the intellectual property involved.
> The don't have to release ALL of it. Releasing just the bits they DO own would be far far better than releasing nothing at all.

Better for who? What does HP get out of giving away their IP? Other than
immense risk!!! Might be better for the denizens of C.O.V but HP, hardly.

> Java - only the bits owned by Sun were released. The rest was replaced by the community (see: OpenJDK)

Don't have any experience with it, but I imagine it is more than an
order of magnitude simpler than VMS.

> Solaris - Again, not all of it was released. What wasn't was replaced by the community (see: Illumos)

Solaris was just another Unix. It had been "replaced" decades ago.
So finding already written pieces to fill in a few blanks was probably
trivial.

> BeOS - only Tracker, the desktop environment, was released IIRC (compiler and text-mode software was all GNU stuff). The rest of the operating system (kernel, drivers, application frameworks, windowing system, graphical utilities, etc) was rewritten by the community (see: Haiku Project)
> Tru64 - only AdvFS was released. AFAIK no one has attempted to rewrite the rest of the OS yet :)

Huh? Tru64 is just another Unix. All of its functionality has been
available in open source for decades.

> Obviously the more HP can release the easier it will be to keep OpenVMS alive. But if they release nothing then OpenVMS probably has no future.

And it still begs the question what gain could HP expect that would
justify the risk.

> So saving OpenVMS probably is possible if HP can be convinced to open source a useful amount of it. Otherwise whatever they do release may at least help people port their software to other platforms.

See above....

> Instead of everyone discussing how hard it is, listing all the reasons why HP might not do it, etc. How about we seriously discuss how we could actually make something happen. What are the options? Has anyone actually tried contacting people at HP? Talking to people who make decisions?

Not regarding VMS but I have been involved in such discussions
with HP on other things. It has always been a non-starter at
the insistence of HP's lawyers who definitely see the risks.

> Remember that HP still has a bunch of people on staff looking after OpenVMS. People who might be able to help. Any attempt to save it is best made bef=

> ore all these people are reassigned.

Good luck.

[Stephen Hoffman]

On 2013-06-13 05:15:31 +0000, JF Mezei said:

> On 13-06-12 21:01, Arne Vajhøj wrote:
>
>> But the chance that there are third party source code in a code base
>> as big of VMS is pretty big.
>
> The DHCP server used to be by a company called "Join". It no longer
> exists. So who would sue HP over such rights ? And what if HP had full
> rights, including open sourcing upon Join ceasing to exist ?

Who really cares about a (missing) DHCP server?

VMS doesn't operate particularly well as a DHCP client — many servers
don't — and no open-sourcing of VMS nor any hypothetical porting of VMS
would ever have a DHCP server anywhere near the top of the
critical-features or porting-priority schedule. It's just too easy to
run with a different DHCP server in a plug computer or a random Linux
or BSD or OS X box, or in a network box, or use Windows Server Active
Directory and its DHCP and DNS services, after all.

Further, good and capable open-source alternatives already exist for
this capability, including the ISC DHCP server
<http://www.isc.org/downloads/dhcp/>, and various other DHCP server
packages. Desperately need a DHCP server prior to its eventual
inclusion in this hypothetical open-sourcing and any subsequent
hypothetical port? Port it yourself.

This all assuming the whole of the TCP/IP Services stack didn't get
overhauled and/or integrated into the base OS in this hypothetical
open-sourcing and any subsequent hypothetical port. Swapping out the
current mail server stuff for Postfix or Lamson, for instance. Not
having to deal with random code that just falls over when pieces are
missing, or having a stack of stuff that sites have to install, or
dealing with all the the chunder-code necessary to "correctly" contend
with and test the error paths for any missing prerequisite products, in
other words.

A DHCP Server is easy to deal with. There are much bigger challenges lurking.


--
Pure Personal Opinion | HoffmanLabs LLC

[John Reagan]

"David Froble" wrote in message news:kpbn9m$27n$1...@dont-email.me...

> Maybe Brian should start sending John Reagan a monthly case of his
> favorite brew.

Who's favorite brew? Brian's or mine? I'm more of red wine drinker. Now,
if somebody wants to send me a case of Merryvale Profile, I won't mind.

[John Reagan]

"JF Mezei" wrote in message
news:51b97b77$0$17011$c3e8da3$76a7...@news.astraweb.com...

> Me thinks the x86 version of Macro64 was done many years ago.
> Wasn't there a Bliss for x86 rumoured done as well ?

Macro64? We were talking Macro32...

There has never been either a Macro64 or Macro32 compiler that generated
Intel x86 code. Never.

There was a BLISS-32 compiler for 32bit mode x86. That was needed to get
GEM and Visual Fortran over to Windows. I believe RDB also was trying to
use this compiler at one point as well. It wasn't a complete implementation
as it only needed to support enough BLISS to compile GEM, etc. It followed
the Windows calling standard, generated Windows objects, Windows debug info,
etc. No BLISS-64. No knowledge of the newer 64-bit registers on x86.
We've discussed this before on the newsgroup.

[Jan-Erik Soderholm]

John Reagan wrote 2013-06-13 16:22:
>
>
> "JF Mezei" wrote in message
> news:51b97b77$0$17011$c3e8da3$76a7...@news.astraweb.com...
>
>> Me thinks the x86 version of Macro64 was done many years ago.
>> Wasn't there a Bliss for x86 rumoured done as well ?
>
> Macro64? We were talking Macro32...
>
> There has never been either a Macro64 or Macro32 compiler that generated
> Intel x86 code. Never.
>
> There was a BLISS-32 compiler for 32bit mode x86. That was needed to get
> GEM and Visual Fortran over to Windows. I believe RDB also was trying to
> use this compiler at one point as well.

It was a working Rdb8 kit but was never released as a fully
supported product due to support issues around Bliss for NT.
There was a ZIP download that I know that I had once,
but I can't find it right now...

Jan-Erik.

[Phillip Helbig---undress to reply]

In article <slrnkrhom...@rayleigh.systella.fr>, JKB
<j...@koenigsberg.invalid> writes:

> > But isn't freeVMS supposed to be a VMS clone? How many years or decades
> > has it been in development? How far along is it?
>
> Sorry, but I cannot understand this argument. On mailing list, I
> have _today_ more than 1000 real subscribers. But all subscribers
> are in read only mode.

OK.

> Until 0.3, we are only _two_ active developers. If there were more
> than two overbooked developers FreeVMS should be usable _today_. But
> I think that all potential users wait for a completed FreeVMS and
> don't want to contribute.

Maybe chicken-and-egg: people won't contribute unless they think it has
a chance of success. How many people developed VMS over how many years?
People doing it part-time might be able to put in, say, 20 hours a week,
so that double that number of people (assuming they are all skilled) and
you still have several years.

[Paul Sture]

In article <b1tujk...@mid.individual.net>,
bi...@server2.cs.uofs.edu (Bill Gunshannon) wrote:

> In article <51b95574$0$63709$c3e8da3$1e92...@news.astraweb.com>,

>
> >
> > If creditors inrehited the IP from Join, wouldn't it be their
> > responsability to knock on HP's door and introduce themselves with:
> > "hey, from now on, you answer to us whenever you need to do something
> > with the Join DHCP server". ??
>
> Not necessarily. They could happily wait around holding on to their
> licenses, wait for someone (like HP) to violate the license and then
> seek monetary damages. Sounds like a good way to milk that cow. And
> it keeps lawyers in work.

Yep. Witness the patent trolls who are happy to buy up IP by the
truckload and see if something juicy comes out of it.

>
> I don't know where it stands but the last I saw about Jstars
> there was talk of moving it all to Windows based systems. Take
> up less space, consume less power, cost considerable less money.

Head for the hills / build a deep cave etc... :-)

--
Paul Sture

[JKB]

Le Thu, 13 Jun 2013 16:29:10 +0000 (UTC),
Phillip Helbig---undress to reply <hel...@astro.multiCLOTHESvax.de> écrivait :

I'm using linux since 1995. In 1995, linux was usable (stable
kernel, X11/openlook, network...). It was a 1.0.9 kernel if I
remember. I don't know how many people have worked on this kernel,
but the first announce concerning a new unix-like kernel was done in
1991. And Linus was alone.

JKB

--
Si votre demande me parvient sur carte perforée, je titiouaillerai très
volontiers une réponse...
=> http://grincheux.de-charybde-en-scylla.fr

[Phillip Helbig---undress to reply]

In article <51b97b77$0$17011$c3e8da3$76a7...@news.astraweb.com>, JF

Mezei <jfmezei...@vaxination.ca> writes:

> Wasn't there a Bliss for x86 rumoured done as well ?

No rumour; it existed so that Rdb could be ported. But nothing came of
that.

> And me thinks Hoff is quietly laughing while reading all this
> speculation from an x86 laptop running VMS natively in his living room.
> Or perhaps he is crying because he is bound to secrecy and can't
> announce that he's already ported VMS.

No smiley; do you really believe that Hoff (even though he knows more
about VMS than both of us together ever will) could actually port VMS
himself?

> And we'll never know because HP isn't going to fess up to the fact that
> the pilot project to evaluate porting VMS to x86 went quite far and
> would be a no brainer to complete if the original engineering team were
> reconstituted.

Reference, please.

> Looks guys, it's over. We can laugh about it. We have to laugh about it.
> Not to rain on your parade but HP has planned this a long time ago, and
> it isn't a small number of grumpy ex VMS customers (hobbyists) who will
> change HP's plans.

I tend to agree here.

[Phillip Helbig---undress to reply]

In article <slrnkrjt1...@rayleigh.systella.fr>, JKB

<j...@koenigsberg.invalid> writes:

> I'm using linux since 1995. In 1995, linux was usable (stable
> kernel, X11/openlook, network...). It was a 1.0.9 kernel if I
> remember. I don't know how many people have worked on this kernel,
> but the first announce concerning a new unix-like kernel was done in
> 1991. And Linus was alone.

Yes, but a) VMS is a lot more complex than a unix-like kernel. Also,
one thing I agree with Stallman about is that it should be called
Gnu/Linux. Much of what people normally think of as Linux are actually
Gnu utilities. Much more time and many more people went into developing
them, and the start was before Linus had his kernel.

[Sprag]

Except that if you look at the contents of a typical Linux distribution, the software from the FSF are a small portion of the overall total. There are tons of non-FSF GNU-, MIT-, Apache-, and Artistic- licensed programs. RMS was pissed that the FSF dropped the ball when it came to building a kernel and started his GNU/Linux crusade to try to draw attention to his cause.

In any case, the biggest challenge to building a VMS replacement is the ecosystem. Linus (and other people who've written their own unix-like kernel) had a huge leg up: the compilers, libraries, and utilities were all available to reuse and/or replace over time. VMS re-implementers don't have that luxury and they've got to start from the ground up.

[Stephen Hoffman]

On 2013-06-13 16:34:03 +0000, Phillip Helbig---undress to reply said:

> In article <51b97b77$0$17011$c3e8da3$76a7...@news.astraweb.com>, JF
> Mezei <jfmezei...@vaxination.ca> writes:
>
>> Wasn't there a Bliss for x86 rumoured done as well ?
>
> No rumour; it existed so that Rdb could be ported. But nothing came of
> that.

"I'll take that secret to my grave, or urn, or medical school
dissecting table or wherever you're dumping me."

>> And me thinks Hoff is quietly laughing while reading all this
>> speculation from an x86 laptop running VMS natively in his living room.
>> Or perhaps he is crying because he is bound to secrecy and can't
>> announce that he's already ported VMS.
>
> No smiley; do you really believe that Hoff (even though he knows more
> about VMS than both of us together ever will) could actually port VMS
> himself?

"Oh suuuure! Last resort.
Old Grandpa the feeb.
The guy who can't be counted on for nothing knowhow dag-nammit.
Everyone's against me!
{Pause}
I'll do it!"

>> And we'll never know because HP isn't going to fess up to the fact that
>> the pilot project to evaluate porting VMS to x86 went quite far and
>> would be a no brainer to complete if the original engineering team were
>> reconstituted.
>
> Reference, please.

I remember some informal lunch-time and hallway conversations, but it's
news to me, too.

[Bill Pechter]

In article <b1s1ra...@mid.individual.net>,
Bill Gunshannon <bill...@cs.uofs.edu> wrote:
>
>Well, many years ago people here were constantly talking about the
>potential good of having VMS still present in academia. But that
>totally aside, do you realize how many universities ran their entire
>enterprise on VMS? Banner was one of the most common administrative
>systems and it ran on VMS. Oracle? Long after academic use had all
>but disappeared here we still had a room full of VMS boxes without
>which the university cold not function. Academic institutions have
>critical systems, too. And even non-profit universities are multi-
>million dollar operations.
>
>bill


Boy... Now I'm totally pissed off. You mean instead of Linux -- Banner ran
on VMS. When did they drop that one? I know they supported mainframes
and AIX...

Just did a net search... they desupported it at Banner 7 which was what
we were running when I got to Seton Hall.

Darn. This place was a total IBM shop when I got here with x86 and
S390... If it had VMS and was going to Linux I'd have been trying to
get the surplus hardware...


Bill

--
--
Digital had it then. Don't you wish you could buy it now!
pechter-at-pechter.dyndns.org http://xkcd.com/705/

[JKB]

Le Thu, 13 Jun 2013 16:44:53 +0000 (UTC),

Phillip Helbig---undress to reply <hel...@astro.multiCLOTHESvax.de> écrivait :

Sure, but in a first time, you can propose VMS API to be only source
compatable. With newlib, you can have a real POSIX subsystem. In
this case, you can reuse a lot of preexistant stuff (like gcc for
example).

[JF Mezei]

On 13-06-13 06:22, David Goodwin wrote:

> HP has no particular reason to scheme away carefully plotting the demise
> of OpenVMS.

Yes they are, from the day they decided to throw in the towel on that
IA64 thing and plan a long term exit from that platform.

> They're a company. Their purpose is to make money.

Unfortunatly, there are people with egos within HP who would not want to
publicly admit that IA64 wa a mistake and killed everythig else just to
save their own image. Then there was the guy who came in whose job was
to cut expenses and decimated in house software development. (probably
because decision to move off IA64 had already been taken, so there was
no point continuing to pretend those operating systems were still being
developped)

> should be seen in that light. If they canceled an x86 port it is because
> they considered it the most profitable (directly or indirectly) thing to
> do.

Again, you underestimate ego. They cancelled all 3 ports because the
HPUX port proved to be too difficult/expensive. It is quite possible
that the VMS port would have been cost effective. But it got taken down
because there was no way HP would port other people's operating systems
and not its own. (VMS has never been "its own" - it was the abstard
adopted child that Digital didn't want and Compaq didn't want)

> Remember who they'd be up against on the x86 platform. Long term even
> Microsoft is going to have a hard time competing with Linux

VMS still has unique capabilities and could remain relevant and
profitable albeit in a niche market if it had continued to be
developped. Note that OS-X still has a small share of the overall
desktop market, but bring in lots of cash to Apple.

> Again, I'd be surprised if there was a plot to kill OpenVMS. They just
> think they have more of a future selling x86 Linux boxen. Nothing
> personal - just business.

If so, then they should come out and announce the formal EOL of VMS and
Itanium. They don't have the guts to do so.

[JF Mezei]

On 13-06-13 12:34, Phillip Helbig---undress to reply wrote:

> No smiley; do you really believe that Hoff (even though he knows more
> about VMS than both of us together ever will) could actually port VMS
> himself?

Yeah, 'cause he's my hero :-) He'd get as far as the DIR command and
then delegate the rest :-)

>> And we'll never know because HP isn't going to fess up to the fact that
>> the pilot project to evaluate porting VMS to x86 went quite far and

> Reference, please.


During the Oracle-HP thing, it was revealed that HP had in fact started
a pilot project to evaluate ports of HP-UX, NSK and VMS to x86. That
project was declared a failure and from that point on, HP started toc
confirm publicly that none of those systems owuld be ported beyond IA64.

[Stephen Hoffman]

On 2013-06-13 18:37:19 +0000, JF Mezei said:

> During the Oracle-HP thing, it was revealed that HP had in fact started
> a pilot project to evaluate ports of HP-UX, NSK and VMS to x86. That
> project was declared a failure and from that point on, HP started toc
> confirm publicly that none of those systems owuld be ported beyond IA64.

HP Project Redwood was the HP-UX port, and got as far as bootstrapping,
but I suspect you know that.

<http://www.theregister.co.uk/2012/05/23/hp_project_blackbird_redwood_hp_ux/page2.html>


If you have citations for any OpenVMS-related x86-64 porting activity,
I'd be interested.

[David Froble]

JF is correct. There is a lot of ego and NIH (not invented here) involved.

Back in the 1990s HP did not believe that out-of-order was going to
work, and so they pursued the IA-64 concept. I seem to recall that back
when IA-64 was trying to get going, and having a hard time of it, that
some people at HP took a look at the success of Alpha and Power and
decided that perhaps they were wrong, and out-of-order needed another
look. The response from HP management was basically "shut up, we know
what we're doing, and it's going to be IA-64".

Now you read things such as Poulson using some out-of order concepts.

IBM knew they were on the right track, and stuck with it. Does anything
today come close to the performance of Power?

Non-technical people at Compaq and HP, being told they were in charge,
assumed that they knew more than the technical people, and made
technical decisions. Alpha was killed off in favor of the sinking
itanic. And soon there will not be a CPU for running VMS and others.

I'm not saying technical people cannot make bad financial decisions, but
for sure, financial people should not try to make technical decisions.
Both would do better without their egos.

[Arne Vajhøj]

On 6/12/2013 11:16 PM, dgso...@gmail.com wrote:
> On Thursday, June 13, 2013 2:41:36 PM UTC+12, Richard B. Gilbert
> wrote:
>> On 6/11/2013 6:21 PM, Single Stage to Orbit wrote:
>>
>>> With the news that HP is to take VMS out the back and put a
>>> bullet through its brains, can we hope that HP will open source
>>> the whole thing so that people can keep using it?
>>
>> I don't believe that they CAN do that. The problem is that H-P
>> does NOT own all the intellectual property involved.
>
> The don't have to release ALL of it. Releasing just the bits they DO
> own would be far far better than releasing nothing at all.
>

> Java - only the bits owned by Sun were released. The rest was
> replaced by the community (see: OpenJDK)

It took SUN 1.5 years from announcement to it was actual available,
which tells a little bit about the required effort.

And the replacements produced by the "community" mostly came
from 1 B$ company (RedHat).

> Instead of everyone discussing how hard it is, listing all the
> reasons why HP might not do it, etc. How about we seriously discuss
> how we could actually make something happen. What are the options?
> Has anyone actually tried contacting people at HP? Talking to people
> who make decisions?

I don't think anyone is trying to prevent anyone from doing so.

Please go ahead.

Arne

[Arne Vajhøj]

On 6/13/2013 2:18 PM, JF Mezei wrote:
> On 13-06-13 06:22, David Goodwin wrote:
>> should be seen in that light. If they canceled an x86 port it is because
>> they considered it the most profitable (directly or indirectly) thing to
>> do.
>
> Again, you underestimate ego. They cancelled all 3 ports because the
> HPUX port proved to be too difficult/expensive. It is quite possible
> that the VMS port would have been cost effective.

Cheaper to port VMS than a Unix to a new ISA? Does not sound likely!

Arne

[Arne Vajhøj]

On 6/13/2013 1:59 AM, David Froble wrote:

> dgso...@gmail.com wrote:
>
>> Instead of everyone discussing how hard it is, listing all the
>> reasons why HP might not do it, etc. How about we seriously discuss
>> how we could actually make something happen. What are the options?
>> Has anyone actually tried contacting people at HP? Talking to people
>> who make decisions?
>

> Yeah, I was thinking about saying something like this. For those who
> want to raise obstacles, go ahead, but some of us just might be able to
> do something.
>
> I don't know that anything CAN be done, but if somebody doesn't try, for
> sure nothing will happen.

True.

But no one is preventing you from starting.

Arne

[Arne Vajhøj]

On 6/13/2013 9:24 AM, Bill Gunshannon wrote:
> In article <1e4252b4-ff19-49d5...@googlegroups.com>,
> dgso...@gmail.com writes:
>> On Thursday, June 13, 2013 2:41:36 PM UTC+12, Richard B. Gilbert wrote:
>>> On 6/11/2013 6:21 PM, Single Stage to Orbit wrote:
>>>
>>>> With the news that HP is to take VMS out the back and put a bullet
>>>> through its brains, can we hope that HP will open source the whole thing
>>>> so that people can keep using it?
>>>
>>> I don't believe that they CAN do that. The problem is that H-P does NOT
>>> own all the intellectual property involved.
>> The don't have to release ALL of it. Releasing just the bits they DO own would be far far better than releasing nothing at all.
>
> Better for who? What does HP get out of giving away their IP? Other than
> immense risk!!! Might be better for the denizens of C.O.V but HP, hardly.
>
>> Java - only the bits owned by Sun were released. The rest was replaced by the community (see: OpenJDK)
>
> Don't have any experience with it, but I imagine it is more than an
> order of magnitude simpler than VMS.

It was around 6 MLOC.

Arne

[Arne Vajhøj]

It could certainly be done.

But try look at what has happened when there has been requests
for help porting some stuff to VMS or to take over maintenance of
something already ported to VMS.

There has not been a long line of people fighting to get the honor.

Arne

[Arne Vajhøj]

But kernel and full OS are different.

Linux got all the userland stuff from other mostly existing projects.

Arne

[Arne Vajhøj]

On 6/13/2013 1:07 PM, Sprag wrote:
> On Thursday, June 13, 2013 12:44:53 PM UTC-4, Phillip
> Helbig---undress to reply wrote:
>> In article <slrnkrjt1...@rayleigh.systella.fr>, JKB
>>
>> <j...@koenigsberg.invalid> writes:
>>
>>> I'm using linux since 1995. In 1995, linux was usable (stable
>>
>>> kernel, X11/openlook, network...). It was a 1.0.9 kernel if I
>>
>>> remember. I don't know how many people have worked on this
>>> kernel,
>>
>>> but the first announce concerning a new unix-like kernel was done
>>> in
>>
>>> 1991. And Linus was alone.
>>
>>
>>
>> Yes, but a) VMS is a lot more complex than a unix-like kernel.
>> Also,
>>
>> one thing I agree with Stallman about is that it should be called
>>
>> Gnu/Linux. Much of what people normally think of as Linux are
>> actually
>>
>> Gnu utilities. Much more time and many more people went into
>> developing
>>
>> them, and the start was before Linus had his kernel.
>
> Except that if you look at the contents of a typical Linux
> distribution, the software from the FSF are a small portion of the
> overall total. There are tons of non-FSF GNU-, MIT-, Apache-, and
> Artistic- licensed programs.

So what?

The fact is that Linux got all the userland stuff from elsewhere.

Whether it was true FSF stuff or non-FSF GNU stuff or non-GNU stuff
is irrelevant for the point.

Arne

[Arne Vajhøj]

On 6/13/2013 2:07 PM, JKB wrote:
> Le Thu, 13 Jun 2013 16:44:53 +0000 (UTC),
> Phillip Helbig---undress to reply <hel...@astro.multiCLOTHESvax.de> écrivait :
>> In article <slrnkrjt1...@rayleigh.systella.fr>, JKB
>> <j...@koenigsberg.invalid> writes:
>>
>>> I'm using linux since 1995. In 1995, linux was usable (stable
>>> kernel, X11/openlook, network...). It was a 1.0.9 kernel if I
>>> remember. I don't know how many people have worked on this kernel,
>>> but the first announce concerning a new unix-like kernel was done in
>>> 1991. And Linus was alone.
>>
>> Yes, but a) VMS is a lot more complex than a unix-like kernel. Also,
>> one thing I agree with Stallman about is that it should be called
>> Gnu/Linux. Much of what people normally think of as Linux are actually
>> Gnu utilities. Much more time and many more people went into developing
>> them, and the start was before Linus had his kernel.
>
> Sure, but in a first time, you can propose VMS API to be only source
> compatable. With newlib, you can have a real POSIX subsystem. In
> this case, you can reuse a lot of preexistant stuff (like gcc for
> example).

But in that case it become a VMS-like OS not VMS.

Arne

[JF Mezei]

On 13-06-13 21:04, Arne Vajhøj wrote:

> Cheaper to port VMS than a Unix to a new ISA? Does not sound likely!

VMS is little endian, fits nicely in x86. VMS has been ported twice
before so good separation of harddware/software.

HP-UX is big endian, so porting it would require a fair amount of work.
Recally that it was also too much work for HP to port TruClusters to HP-UX.

It seems to me like HP-UX may not the best code base in terms of
structure and flexibility.

Personally, I would not be surprised to see NSK ported to x86. It is a
low key OS used by trulty mission critical sites willing to pay the big
bucks. They really do need the redudancy etc because downtime is
measured in lives or millions of dollars.

[JKB]

Le Thu, 13 Jun 2013 21:10:53 -0400,
Arne Vajhøj <ar...@vajhoej.dk> écrivait :

I know. And ?

[JKB]

Le Thu, 13 Jun 2013 21:15:56 -0400,
Arne Vajhøj <ar...@vajhoej.dk> écrivait :

Maybe you prefer a dead OpenVMS than an alive clone of OpenVMS.
Don't forget that Linux (or xBSD) is often better than some
proprietary Unix.

[Michael Kraemer]

JF Mezei schrieb:

> Recally that it was also too much work for HP to port TruClusters to HP-UX.

The way more probable reason was that HP-UX' standard JFS filesystem
was done by Veritas, so it was just natural (and cheaper) to let
them do the work, rather than trying to integrate an alien like AdvFS.

> It seems to me like HP-UX may not the best code base in terms of
> structure and flexibility.

It's an old dog, but so is VMS.
The main difference is customer base, not code base.

[Simon Clubley]

On 2013-06-13, JF Mezei <jfmezei...@vaxination.ca> wrote:
> On 13-06-13 21:04, Arne Vajhøj wrote:
>
>> Cheaper to port VMS than a Unix to a new ISA? Does not sound likely!
>
> VMS is little endian, fits nicely in x86. VMS has been ported twice
> before so good separation of harddware/software.
>

VMS most certainly does _NOT_ have good separation, at least when compared
the alternatives such as Linux. Take a look at the Linux kernel code base
sometime and see just how cleanly the architecture specific part of the
code is isolated from the generic parts of the code. Now _that's_ clean
separation and is the reason why it's available on so many architectures.

The VMS internal architecture also requires much more hardware support than
Linux does and you have to either emulate that somehow (if that's even
possible) or rewrite the OS if the required hardware support is not present.

For example, you will not see VMS, as it stands, running on ARM because ARM
simply does not the hardware support for the multiple protection rings
VMS requires. The FreeVMS people, by using a microkernel architecture,
have managed to work around this but that requires a VMS rewrite.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world

[Bob Koehler]

In article <b1tvh5...@mid.individual.net>, bi...@server2.cs.uofs.edu (Bill Gunshannon) writes:
>
> Huh? Tru64 is just another Unix. All of its functionality has been
> available in open source for decades.

No. Tru64 had a VMS-like DLM. It's been less than decades since
other UNIX got that, and I'm not sure all of them have it yet.

And Tru64 was much easier to admin than Solaris or HP-UX. So is
AIX.

[Bill Gunshannon]

In article <slrnkrld6...@rayleigh.systella.fr>,

JKB <j...@koenigsberg.invalid> writes:
> Le Thu, 13 Jun 2013 21:15:56 -0400,
> Arne Vajhøj <ar...@vajhoej.dk> écrivait :
>> On 6/13/2013 2:07 PM, JKB wrote:
>>> Le Thu, 13 Jun 2013 16:44:53 +0000 (UTC),
>>> Phillip Helbig---undress to reply <hel...@astro.multiCLOTHESvax.de> écrivait :
>>>> In article <slrnkrjt1...@rayleigh.systella.fr>, JKB
>>>> <j...@koenigsberg.invalid> writes:
>>>>
>>>>> I'm using linux since 1995. In 1995, linux was usable (stable
>>>>> kernel, X11/openlook, network...). It was a 1.0.9 kernel if I
>>>>> remember. I don't know how many people have worked on this kernel,
>>>>> but the first announce concerning a new unix-like kernel was done in
>>>>> 1991. And Linus was alone.
>>>>
>>>> Yes, but a) VMS is a lot more complex than a unix-like kernel. Also,
>>>> one thing I agree with Stallman about is that it should be called
>>>> Gnu/Linux. Much of what people normally think of as Linux are actually
>>>> Gnu utilities. Much more time and many more people went into developing
>>>> them, and the start was before Linus had his kernel.
>>>
>>> Sure, but in a first time, you can propose VMS API to be only source
>>> compatable. With newlib, you can have a real POSIX subsystem. In
>>> this case, you can reuse a lot of preexistant stuff (like gcc for
>>> example).
>>
>> But in that case it become a VMS-like OS not VMS.
>
> Maybe you prefer a dead OpenVMS than an alive clone of OpenVMS.

It's dead OpenVMS in either case. If what people here are lamenting
is the loss of VMS, it really doesn't matter because both are the same
in that respect.

> Don't forget that Linux (or xBSD) is often better than some
> proprietary Unix.

In what way? How is Linux in any way better than AIX? (realizing that
as far as I know AIX and HP-UX are about the last proprietary Unixes
still in active development.)

bill


--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bill...@cs.scranton.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>

[JKB]

Le 14 Jun 2013 13:47:00 GMT,
Bill Gunshannon <bi...@server3.cs.uofs.edu> écrivait :

> In article <slrnkrld6...@rayleigh.systella.fr>,
> JKB <j...@koenigsberg.invalid> writes:
>> Le Thu, 13 Jun 2013 21:15:56 -0400,

>> Arne Vajhøj <ar...@vajhoej.dk> écrivait :

>>> On 6/13/2013 2:07 PM, JKB wrote:
>>>> Le Thu, 13 Jun 2013 16:44:53 +0000 (UTC),

>>>> Phillip Helbig---undress to reply <hel...@astro.multiCLOTHESvax.de> écrivait :

You have forgotten Solaris. Linux (and all BSD) is better than all
other Unix as you cas install and easly port all programs from an
architecture to a new one. Solaris, AIX, HP-UX have certainly a lot
of features that Linux hasn't, but Linux (and all open source
operating system) gives portability.

[Bill Gunshannon]

In article <slrnkrm8b...@rayleigh.systella.fr>,

No, I haven't. It is not being actively developed (on the proprietary
side) as much as AIX and even HP-UX are. And the gap between the open
source and proprietary versions is very murky.

> Linux (and all BSD) is better than all
> other Unix as you cas install and easly port all programs from an
> architecture to a new one.

Ridiculous. Linux programs frequently don't port easily to BSD and BSD
programs frequently don't port easily to Linux. About that same as going
between either of them and AIX or Hp-UX (or Solaris if including them
makes you happy!)

> Solaris, AIX, HP-UX have certainly a lot
> of features that Linux hasn't, but Linux (and all open source
> operating system) gives portability.

Hogwash and I have been running and porting between Linux, BSD
and pretty much all of the proprietary Unixes for about 30 years
now. Heck, I have run into programs in the past that wouldn't
port easily between different flavor of Linux!! The portability
of the application is much more dependant on the writer of the
application than the OS. If you stick to common API's and don't
use non-portable tricks they are all pretty much the same.

I have even run into applications that were not portable between
32 and 64 bit versions of the same release of Linux and BSD. I'll
leave it to you to figure out why that might be. :-)

[JKB]

Le 14 Jun 2013 14:15:07 GMT,

Bill Gunshannon <bi...@server3.cs.uofs.edu> écrivait :
> In article <slrnkrm8b...@rayleigh.systella.fr>,
> JKB <j...@koenigsberg.invalid> writes:
>> Le 14 Jun 2013 13:47:00 GMT,

>> Bill Gunshannon <bi...@server3.cs.uofs.edu> écrivait :

>>> In article <slrnkrld6...@rayleigh.systella.fr>,
>>> JKB <j...@koenigsberg.invalid> writes:
>>>> Le Thu, 13 Jun 2013 21:15:56 -0400,

>>>> Arne Vajhøj <ar...@vajhoej.dk> écrivait :

>>>>> On 6/13/2013 2:07 PM, JKB wrote:
>>>>>> Le Thu, 13 Jun 2013 16:44:53 +0000 (UTC),

>>>>>> Phillip Helbig---undress to reply <hel...@astro.multiCLOTHESvax.de> écrivait :

Not ridiculous. If you have written a program for Linux/Alpha, you
can recompile this application without modification on Linux/amd64
or Mips.

>> Solaris, AIX, HP-UX have certainly a lot
>> of features that Linux hasn't, but Linux (and all open source
>> operating system) gives portability.
>
> Hogwash and I have been running and porting between Linux, BSD
> and pretty much all of the proprietary Unixes for about 30 years
> now. Heck, I have run into programs in the past that wouldn't
> port easily between different flavor of Linux!! The portability
> of the application is much more dependant on the writer of the
> application than the OS. If you stick to common API's and don't
> use non-portable tricks they are all pretty much the same.
>
> I have even run into applications that were not portable between
> 32 and 64 bit versions of the same release of Linux and BSD. I'll
> leave it to you to figure out why that might be. :-)

I write applications that have to be portable on i386, amd64, mips,
arm (32 and 64) and sparc (32 and 64). If you correctly write your
application, it is portable without effort.

[Bill Gunshannon]

In article <slrnkrma0...@rayleigh.systella.fr>,

Not always. Try writting a program that makes assumptions about byte
order or individual data item size. then try moving it around between
different machines.

>
>>> Solaris, AIX, HP-UX have certainly a lot
>>> of features that Linux hasn't, but Linux (and all open source
>>> operating system) gives portability.
>>
>> Hogwash and I have been running and porting between Linux, BSD
>> and pretty much all of the proprietary Unixes for about 30 years
>> now. Heck, I have run into programs in the past that wouldn't
>> port easily between different flavor of Linux!! The portability
>> of the application is much more dependant on the writer of the
>> application than the OS. If you stick to common API's and don't
>> use non-portable tricks they are all pretty much the same.
>>
>> I have even run into applications that were not portable between
>> 32 and 64 bit versions of the same release of Linux and BSD. I'll
>> leave it to you to figure out why that might be. :-)
>
> I write applications that have to be portable on i386, amd64, mips,
> arm (32 and 64) and sparc (32 and 64). If you correctly write your
> application, it is portable without effort.
>

And that's the crux. It is not the OS that makes it portable. It
is the effort of the programmer. The same is true of proprietary
Unixes. If you write your program using API's that are not common
among all the machines you want to run it on it will not be portable
and vice versa if you stick to API's that are common the program will
port easily to any Unix. Ever see a program that took in characters
and then tested to see what key was typed by looking at the numeric
value (ie. "A" == 65)? Try running that program under Primix. :-)
It is a lot easier to wtrite non-portable programs than portable ones.
Most people opt for the easy way out. And then blame the system.

[JKB]

Le 14 Jun 2013 15:07:40 GMT,

Bill Gunshannon <bi...@server3.cs.uofs.edu> écrivait :
> In article <slrnkrma0...@rayleigh.systella.fr>,
> JKB <j...@koenigsberg.invalid> writes:
>> Le 14 Jun 2013 14:15:07 GMT,

>> Bill Gunshannon <bi...@server3.cs.uofs.edu> écrivait :

>>> In article <slrnkrm8b...@rayleigh.systella.fr>,
>>> JKB <j...@koenigsberg.invalid> writes:
>>>> Le 14 Jun 2013 13:47:00 GMT,

>>>> Bill Gunshannon <bi...@server3.cs.uofs.edu> écrivait :

>>>>> In article <slrnkrld6...@rayleigh.systella.fr>,
>>>>> JKB <j...@koenigsberg.invalid> writes:
>>>>>> Le Thu, 13 Jun 2013 21:15:56 -0400,

>>>>>> Arne Vajhøj <ar...@vajhoej.dk> écrivait :

>>>>>>> On 6/13/2013 2:07 PM, JKB wrote:
>>>>>>>> Le Thu, 13 Jun 2013 16:44:53 +0000 (UTC),

>>>>>>>> Phillip Helbig---undress to reply <hel...@astro.multiCLOTHESvax.de> écrivait :

I agree, but in this case, it's a bug in your code.

> The same is true of proprietary
> Unixes. If you write your program using API's that are not common
> among all the machines you want to run it on it will not be portable
> and vice versa if you stick to API's that are common the program will
> port easily to any Unix.

If I remember, I have written you have to restrict your usage to
POSIX, or SysV or BSD API. If you try to use anonymous semaphores on
Mac OS X, or shm_open() on NetBSD, sure, you will have some strange
portability issues.

> Ever see a program that took in characters
> and then tested to see what key was typed by looking at the numeric
> value (ie. "A" == 65)? Try running that program under Primix. :-)

And ? "A" == 65 is not portable. I can write a piece of code that is
not portable anywhere. IOCCC is full of this piece of code.

> It is a lot easier to wtrite non-portable programs than portable ones.
> Most people opt for the easy way out. And then blame the system.

[Johnny Billquist]

On 2013-06-12 23:59, David Froble wrote:
> Stephen Hoffman wrote:
>> On 2013-06-12 19:43:34 +0000, John Wallace said:
>>
>>> OK, the principle is understood, but are you (or is anybody) aware of
>>> specific relevant examples where it applies (or even might apply) to
>>> real VMS source code?
>>>
>>> There are people round here with VMS listings. I'm not one of them,
>>> but surely if it's been made obvious in the listings that there are
>>> restrictions on the IP, someone might know by now....
>>>
>>> Or have I missed something (it happens)?
>>
>> The source listings are expurgated. Not all of the code of VMS is
>> included in those kits.
>
> Ok, here's a question. Could it be assumed that the source listings are
> free of any IP and anything else that HP could not "open source" or
> otherwise make available? If so, then question #2 becomes, how much of
> VMS is there? Would it be a good start on a non-HP VMS?

Assume nothing, when it comes to intellectual property. *Everything*
will need to be vetted. Absolutely nothing will come out before it has
been audited. Which implies a big process that there are chances you
will not see HP paying for.

Johnny

[Johnny Billquist]

On 2013-06-14 04:39, JF Mezei wrote:
> On 13-06-13 21:04, Arne Vajhøj wrote:
>
>> Cheaper to port VMS than a Unix to a new ISA? Does not sound likely!
>
> VMS is little endian, fits nicely in x86. VMS has been ported twice
> before so good separation of harddware/software.

What have you been smoking?

> HP-UX is big endian, so porting it would require a fair amount of work.
> Recally that it was also too much work for HP to port TruClusters to HP-UX.

Endianness is not the biggest problem around. That said, I don't know
what odd things might sit in HP-UX, but any Unix is almost by default
way easier to port than VMS.

Johnny

[John Wallace]

On Jun 14, 7:59 am, Simon Clubley <clubley@remove_me.eisner.decus.org-
Earth.UFP> wrote:

Heard this a couple of times recently. I know how many modes a VAX
has, I know how many "protection rings" an x86 offers in principle,
but what number is applicable to the Alpha architecture, and where can
one find a relevant description (doesn't seem to be obbvious in the
freely downloadable Alpha Architecture Handbook. might be in the
Architecture Reference Manual but I'm away from mine, might be in the
DTJ somewhere)?