[Info-vax] date comparison format from a program

[gérard Calliet]

Le 21/05/2019 à 16:43, Dave Froble a écrit :
> On 5/21/2019 1:14 AM, gérard Calliet wrote:
>> Le 21/05/2019 à 04:15, Dave Froble a écrit :
>>> On 5/20/2019 6:44 PM, Hans Bachner wrote:
>>>> gérard Calliet schrieb am 20.05.2019 um 18:44:
>>>>> [...]
>>>>> Yes, but they are totally isolated platforms, and everything (software
>>>>> and hardware) on it has been certified at the time of the development.
>>>>>
>>>>> They have now to garanty a maintenance in operational condition.
>>>>> Reverse
>>>>> engineering on VAX CPU board, validation of parts are the major work.
>>>>> [...]
>>>>
>>>> Do they want to build spare CPU boards?
>>>>
>>>> Actually, reverse engineering of various VAX main boards has already
>>>> been done, but the result is software... check out CHARON-VAX, whether
>>>> it fulfills your certification requirements.
>>>>
>>>> At least, the software passed the original DEC hardware diagnostic
>>>> routines.
>>>>
>>>> Hope this helps,
>>>> Hans.
>>>
>>> I think the key word here is "certification".
>>>
>>> Someone has certified particular HW, running a particular version of
>>> VMS, and some application(s).
>>>
>> It's exactly the case.
>>> Doesn't do any good to mention out of date OS, old HW, VAX C,
>>> emulators, security, and just about anything else.
>> Right, anything new is impossible.
> 
> Actually, that is a choice, not reality.  And it's a choice that I do 
> not agree with.  Not that that matters.
>
That has been a choice when it was developped, but now it works in 
extremely sensitive environments, all elements have ben certified with 
very long certifications. So now the users decided there will be not any 
change.

> Now, I may be wrong, same old territory for me, but, if I was designing 
> something that had to last for a long time, I'd design it with "black 
> boxes" that perform a task, but can be replaced (inside the box) with 
> newer stuff.  As long as the "box" did what it was suppose to do, so 
> what?  More viable, probably cheaper.
Developed in 1990s, there were a lot less options.

And "black boxes" could be right. But major bugs now come with black 
boxes which have not-too-specified side effects, and often ways for big 
problems. Black boxes with not any side effect and Qualities of Service 
totally garantied. Yes. You know some?
> 
> Yeah, there can be issues, but losing some irreplaceable component is 
> also an issue.
> 
>>> Building VAX CPU boards is really way out there.  But, if it's the
>>> job, then it's the job.
>> They tried that in the beginnings (1990s), thinking just replacing parts
>> in CPU boards could help sometime... And it was not worth it: these CPU
>> boards don't bug for decades.
> 
> Tell me about it.  VAXstation 4000 stuff is what, 1990?  That's almost 
> 30 years.  Still running here.
VAX 4000 500. With a special board developped just for the customer (and 
now offered by nemonix) to have SCSI disks.
> 
> The N-VAX chip could have evolved into a system-on-a-chip and been dirt 
> cheap to produce in the millions.  That would have been plenty good for 
> many things.  Better than that x86 POS.
> 
>