[Info-vax] Somewhat levels up to not port outside VMS (letter from Wyoming)

[gérard Calliet]

>> The problem is that you express what you say using this over the top
>> flowery and poetry based language instead of just coming out and
>> saying directly exactly what you mean.
> 
I’ll try to adapt to c.o.v.

I have been wholly upseted here by an advice from Stephen Hoffman who 
adviced to propose to the customer a port outside VMS. And more than 
that I understood I had to level up for that. And I began a too 
passionated argument. Apologies for that.

I do appreciate here the huge encyclopedic knowledge, and Stephen 
Hoffman is not the least encyclopedist.  But also I do think « 
comparaison n’est pas raison » (to compare in not to think). If we do 
just comparisons to decide if we have to stay or not on VMS, perhaps 
Stephen could be right in a lot of circonstances. But the reasons why it 
is good to stay on VMS are for me far beyond the horizontal comparisons.

I tried to give some of the fundamental reasons who are behind my choice 
to stay on VMS. I try here to be more simple.

First: use your brain: I gave a riddle with all the epigraps of the Ruth 
Goldenberg book on VMS internals. The time where VMS has been made was a 
time where computer scientists did love the multiple usages of brain. 
The epigraphs prove it. And I’m very proud that one of the good answers 
were by Camiel. It is a sign that the new lovers of VMS like it 
*because* they see it as a good brain usage. And thanks to Craig, who 
knows that.

Second: Bayes theorem and commen sense. A lot of things are done now on 
Artificial Intelligence thanks to the Bayes theorem on conditional 
probabilities. It seems our brain works a little bit like that theorem. 
And this is as simple as « you go faster and accuracier if you have got 
experience ». As simple as that.

Third: Idiom. The best application developped on VMS have used the best 
specific structures available on VMS. It is their idiom. Everyone in 
linguistics explain that idiomatic expression are a nightmare to the 
translator. Why searching to translate when translation is more than 
difficult? And when the used idiom is a good structured langage?

Fourth:  Signs. A good application provides to the user a representation 
of the job. Sometimes I agree it could be nice to change. And I know 
Word users who do like to change the way they edit a text each sixt 
months, because M. Microsoft decided to change Word. But in the case 
that the interface, for example, is one of an important tool on an 
urgency resuscitation service, I doubt changing every six months the 
representation of the bodies is a good choice.

Five: Proximity. VMS developped as a departmental offer (neither main 
frame, nor desk computer). It is perhaps the exact opposite of cloud 
computing paradigm. Cloud computing can be very usefull for a lot of 
applications, and very ugly for others. VMS will help a lot for 
application of proximity, for example for control of factories.

Six: Predictable behavior against Murphy law. With the huge ressources 
we can get safety with probalistics methods, and with the big data, we 
can get good predictions by correlative methods.  But « Tout ce qui est 
susceptible de mal tourner tournera mal, et corrélation n’est pas raison 
». In certain cases we have to be able to get ully predictable 
behaviors, depending on rational reasoning, provable by formal methods. 
And in these contexts it is better to know why and when something (even 
bad) can happen than to pray that the Murphy law will not be verified.

I did my best to be the simpler I can be. (As adviced by Dave and Craig, 
thanks).

I am aware these six reasons are a lot abstract, and I don’t pretend it 
could be business efficient to give them to a customer. On the other 
side giving only the technical reasons we are all knowing, including the 
50 points developped by Eddie Orcutt, we are exposed to the good counter 
arguments of Stephen Hoffman. Something has to be invented between pure 
technical and the totally abstract.

I was thinking c.o.v. could be a place where this type of work could be 
done. I thought c.o.v. was *for* VMS and not only *about* VMS. My mistake?

For now I have just given to Stephen Hoffman the reasons why I don’t 
propose porting out VMS. And even as a tactic, as said by Dave, I’m not 
sure it is a good idea.

If any of my 6 abstract reasons inspire discussion, it would be a 
pleasure to go on, and I'm waiting for any other good ideas to promote VMS.

Gérard Calliet