[Info-vax] Interesting View of Graphics Driver Futures

[Stephen Hoffman]

On 2016-06-16 04:15:01 +0000, lawrencedo99 at gmail.com said:

> On Saturday, May 7, 2016 at 3:00:21 PM UTC+12, IanD wrote:
>> I think it was about at this time I also realised how disjointed linux 
>> support was. Hundreds to 1000's of posts to go over, searching for that 
>> elusive post where someone had the same card, same system and config 
>> and then could communicate without being so terse that only a linux 
>> guru could understand

Ayup.   But also try doing some of that same configuration and 
controller research with OpenVMS.   Figuring out what's supported on 
OpenVMS isn't getting any easy, either.  And the support matrix is a 
lot smaller than what Linux supports.   The SPDs omit more than a 
little, SPOCK AFAICT isn't being kept particularly current and looks to 
be headed for replacement, and the QuickSpecs don't always get updated 
for changes in OS support, and the HPE search engine seemingly couldn't 
find itself if you spotted it the URL.   Which usually then means 
direct contact with the escalation technical support and engineering 
teams, when you're trying to figure out if specific (for instance) HPE 
storage hardware is supported with HPE OpenVMS.

>> This is what set VMS support apart imo. Helpful and clearly articulated 
>> and with actual examples
> 
> Question of scaling, I guess: Linux has to deal with a several orders 
> of magnitude greater combination of possible hardware configurations 
> than VMS does.

Performance scaling, too.  OpenVMS was advertised with the "desktop to 
data center" tag line, but that range has increased substantially in 
recent years.  From mobile supplanting more than a little of the 
desktop use to far larger data centers orders of magnitude larger than 
clusters can provide, OpenVMS hasn't kept up.


-- 
Pure Personal Opinion | HoffmanLabs LLC