[Info-vax] relaunch or legacy

[Dave Froble]

On 1/28/2022 7:58 AM, Gérard Calliet wrote:
> Le 28/01/2022 à 13:29, John Dallman a écrit :
>> In article <j5i0m8Fs463U1 at mid.individual.net>,
>> gerard.calliet at pia-sofer.fr (Gérard Calliet) wrote:
>>
>>> I apologize for the probably pretentiousness of this presentation.
>>> If it can be an excuse, I consider that the respect due to an
>>> audience as learned and experienced as c.o.v. implies to entrust
>>> them with complex things.

Frankly, I fail to understand just what you're trying to say.  Maybe use fewer 
and smaller words?

>> Here we seem to have a difference in national etiquette. To Americans,
>> and other English-speakers, especially engineers, a crisis is best
>> explained in a few blunt words. Lengthy speeches remove the sense of
>> urgency.

Just what is your crisis?  Try fewer words.  Be specific.

>> However, I don't think the situation is as bad as you do.
> I cannot say anything about the world wide - because VSI don't say anything -.
> But in France the situation is a lot worse than what I say. And we have had in
> France a lot of efforts made to keep the customers in (port of python, use on
> vms of zabbix,... a heir of DECUS doing a lot of things (100 attendees on
> meetings about VMS with VMSgenerations). And in the same country I think I'm the
> only one who says VSI can succeed. Others (customers and consultants)all say VMS
> will dye in 3 or 5 years. The differences are only between angst and anger.
> VSI have not

If there is such an attitude, that VSI will fail, what are the facts behind such 
an attitude?

VSI is rather far along the road to success.

>> made an elaborate plan to address all of the things you're worried about,
>> but that is almost certainly because they've been concentrating on the
>> issue of the x86 port.
> My point. On my side I see that as the bad idea.

The x86 port is the reason VSI exists.  I'm not sure what else you want them to 
do.  Can you be specific?

> I agree without x86 in some
> future, nothing is possible. But because the time to x86 is long - more long
> every year - it is important to cope with a very long transition, and so the
> unique goal x86 cannot work. And even if we had now x86, a port is always a big
> decision(remember the ports to alpha or itannium),

Yes, I remember, and they happened rather easily.  For many, it was compile, 
link, and run.  Well, yeah, moving data to the new system too, but that's not 
really part of any port.

Now, if you don't have source code, better look for emulators, because you're 
screwed.

> we have the problem of ISV.
> The logic had to be we will port to x86 because VMS is good for us for x, y, z
> questions, notably the confort that give us VSI we'll try a port to x86. And
> not: because you will have x86 you have to be with VMS now, even with
> sacrifices. It is this logic which doesn'nt work.

That made no sense to me ...

> Without that working, they are sunk. Now they know
>> it will work, they should be making the plans for distribution and
>> marketing.
> No. Marketing had to be and has to be made about VMS intrinsics. Same idea.>
>> Remember that we're viewing this process partly from the inside. It is
>> not surprising that it does not look smooth; these things never do from
>> within.
>>
>>> I come back to the "Digital is dead, long live DEC" reference. I
>>> always have this reference in mind, following two ways of thinking:
>>
>> I'm afraid that to me it does not convey anything meaningful.

Correct ...

>> I can, however, see a possible approach that would assist French
>> customers and intermediaries. Let me explain:
>>
>> A transition to x86 is excellent for VMS end-users who are willing and
>> able to move their applications to x86 swiftly. But not all of them are
>> in that position.

Why not?

>> A transition to x86 means that some parts of intermediaries' expertise
>> becomes far less useful: expertise on DEC and the older HP hardware will
>> no longer be required by end-users who are no longer using old kit. The
>> intermediaries need something else to sell.

That really is not a VMS problem.  Perhaps a business problem.

>> These two problems may have the same solution.
>>
>> I mentioned the idea of emulating Alpha and IPF yesterday. Emulating
>> 64-bit VMS on VMS is rather easier than emulating it on a non-VMS
>> operating system. The x86 VMS has, I think, all the system calls of Alpha
>> and IPF versions, and thanks to the DEC calling standard, they're called
>> in the same way.
>>
>> So reviving the free and open-source IPF emulator I posted a link to
>> today, porting it to x86 VMS, and equipping it with a system call
>> translation facility would seem to be a route to running IPF VMS
>> applications on x86 VMS.
>>
>> That gives a transition route to running on x86 for IPF customers who
>> can't or won't port their applications, and gives intermediaries
>> something to provide expertise on. This will require a fair amount of
>> open-source software development work, but that will benefit everyone
>> involved, and probably prompt improvements to x86 VMS development tools.
> You gave excellent technical ideas. I'll see that.
>
> The point today is business success, gaining again trust from customers. And
> they are not stupid: we have to answer objective things to their abjective angsts.

The bottom line, for me, is "will VMS on x86 meet your needs".  If so, then use 
it.  If not, then you will be looking elsewhere.


-- 
David Froble                       Tel: 724-529-0450
Dave Froble Enterprises, Inc.      E-Mail: davef at tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA  15486