[Info-vax] inertia or fundamentals about langages?

[gérard Calliet]

Le 22/05/2019 à 18:37, Stephen Hoffman a écrit :
> On 2019-05-22 09:37:48 +0000, gérard Calliet said:
> 
>> 1) you report the un-cleverness to the context. Because if they had 
>> choice, they would not be in inertia. So, or there is not any context 
>> which is free to decide, or the context is in inertia.
> 
> A site that's locked at V5.5-2H4 and K&R VAX C or any other HPE OpenVMS 
> release, and for whatever good or bad reason—or anything related to that 
> Canadian PDP-11 hardware for that matter—are no more the future of VSI 
> and of OpenVMS than punched cards and card readers are the future over 
> at IBM.
> 
>> 2) my point is not to say there is not inertia, I agree with you, 
>> there is such thing, and as you describe it ; but I don't agree with 
>> you if you say that it is the *only* cause of lasting on VMS ; you can 
>> replace "cleverness" by "value", and my opinion is that we stay 
>> (sometimes) on VMS not only because of our inertia or of the inertia 
>> we undergo, but because VMS, by its specific values and by the 
>> specific story of building our solution on it has more value than 
>> other environments.
> 
> VSI and its predecessors have not been able to communicate that value. 
> That value has variously been offset by other difficulties.  And that 
> value has been offset by competitors.
> 
> The results of the last twenty years' OpenVMS upgrades hasn't been 
> entirely competitive.
> 
> Yes, there are and have been and will be sites that fully use OpenVMS 
> for its features and capabilities.
> 
>> 3) the (2) point is decisive to think about a future: it is only if 
>> there is more value on VMS that VMS could get future success ; and yes 
>> it's a commercial product, and its intrinsic value is not the only 
>> cause for a success ; but if we forget the root, which is the value, 
>> perhaps wwe'll not see in anytime the tree.
> 
> The root of the value for VSI for now is that it's OpenVMS, and it's a 
> pain to port many of the existing apps off of OpenVMS.  And it's easier 
> for those folks to upgrade to VSI OpenVMS, if (when?) they need support.
> 
> The root of the value of OpenVMS for end-users?  That varies. For a 
> number of sites, it's that it's a pain to port apps off of OpenVMS. For 
> others, specific OpenVMS features.  That varies.
> 
> No small part of that value for existing OpenVMS sites is the 
> familiarity of the staff.  That's also not a benefit that's easily 
> and/or inexpensively realized by new installations.  That's the "you 
> need trained staff" cost.  You do.  But not for some of the things that 
> OpenVMS needs trained and experienced staff for.  The user interface and 
> API—I recently heard a rather long rant from somebody trying to use 
> OpenVMS to do some simple operations—is a ongoing problem for new 
> adoption.  Took me a few minutes to solve what had stymied them, but 
> that for a situation that should not even exist.  There's a lot of this 
> in OpenVMS.  Too much.
> 
> As for "future success", I've suggested two features that are still 
> somewhat unique.  One—clustering—is priced out of common usage.  The 
> other—multi-host RAID-1—is still fairly unique.
> 
> These and other benefits have clearly been insufficient to draw interest.
> 
> There are lower-end alternatives to what clustering provides many 
> (though not all) apps, with the use NAS and redundant NAS via file 
> shares, and with higher-end and higher-scale clustering configurations 
> up where OpenVMS can't reach.
> 
> And as for clustering, there's that servers have gotten both much more 
> dense and much faster and much more reliable than those of the VAX era. 
> And failover via VM or via database replication or other approaches 
> works pretty well for some folks.
> 
> There are other OpenVMS features sometimes cited—logical names—that have 
> equivalent of better alternatives on other platforms.  An integrated 
> DLM, and that exists in other platforms.
> 
> VSI is working to remove some of the impediments, with the port being 
> one of the earlier and larger impediments being resolved, and there are 
> others.
> 
>>> And folks that are still running V5.5-2H4—clever or not, locally 
>>> appropriate or not—are not part of the future of VSI OpenVMS.
>> Do you really think my only customer is this strange one, and that all 
>> I say is referred to it?
> 
> I know not which you believe strange.
> 
> 
> 
To discuss needs a little bit of honesty. You get your arguments about 
the fact I spoke about a vms 5.5 case. I had to say my analysis around 
vms is not based on this specific case.

For this discussion I think there are a difference of point of vue. My 
opinion here is althought often we can say "the tree hides the forest", 
I hear your analyzis as "a forest which hide the tree".

A lot of (important) details sometimes hampers consideration on 
fondamentals. The revival of VMS is far from a  set of details.

And a lot of contextual causes doesn't everytime prevent success of 
cleverness. The VMS x86 port is one of these stories.

I remember how "poetic" was thought here my analysis on a potential 
revival of VMS - and there was a lot of contextual good reasons to think 
like that -. And it seems it has been my fundamentals (poetics) which 
has been right.

Your analysis are very usefull and always full of precious information. 
Try also synthesis.