[Info-vax] OpenVMS servers and clusters as a cloud service

[Stephen Hoffman]

On 2018-01-10 05:54:50 +0000, Grant Taylor said:

> On 01/09/2018 11:11 AM, Scott Dorsey wrote:
>> There's no reason not to do it, other than coding time.
> 
> The biggest reason I've seen, is one word, "ignorance".  Ignorance of 
> <wildcard>.  (I've seen more than I want to admit to.)

"Blame the user".

We are in the world where an increasing number of folks do not have the 
knowledge or do not the time to acquire and develop that knowledge, or 
lack the schedule time or the budget to create more robust 
applications, and more secure configurations.

We are in the world where implementing robust solutions commonly does 
not happen.

We are in the world where long-running processes routinely do not checkpoint.

Now...

Think hard here, all you fine readers.  Think.  Really think.

What will the best response to this sorry state of affairs be?   If 
it's "blame the users" or "blame the developers" or "I could have done 
that with first principles", look around.  Yes, that might be an 
answer, but is it a viable answer?   Is it a good answer?   Is it an 
answer that any of us can really work with?  Or is it a cop-out, or a 
makes-me-feel-superior, or a palliative?  Is it an answer that's part 
of something that an existing or new product can use in its designs and 
marketing?

Sure, ignoring the problem and shifting the blame onto the end-users or 
the ISVs or the system managers can work for some incumbent businesses. 
 For a while.   That doesn't work for businesses that want or need to 
expand their customer bases.  It doesn't work for the rest of us 
either, as our data and our systems and our servers are increasingly 
interconnected, and increasingly exposed.   Arcane skills — such as 
cryptography — become increasingly central to all of our processing.  
Where a developer mistake with cryptography or a developer that hasn't 
kept current with what's known and available, or a patch is delayed or 
ignored or omitted, can get really expensive.

If providing more robust processing doesn't get easier and simpler and 
more capable on OpenVMS, then it's not going to happen with OpenVMS.  
Because it hasn't happened.  And it's not now going to suddenly start 
happen.  And because potential new customers are either going to 
continue to buy what they already have and are familiar with — which is 
not OpenVMS — or they're going to look elsewhere for their options and 
alternatives, it's not something that VSI can use to differentiate.   
Worse, iIf your current or new product — VSI OpenVMS or otherwise — 
requires more training, or acquiring new skills, or porting efforts, or 
changing tools or moving to what are lesser development and management 
and other tools, and/or you can't easily host instances in common 
places, then you're already operating at a competitive disadvantage.

Going first-principles or blame-the-user in your product approach or in 
your API designs?  That's unlikely to be the basis of any successful 
product development or product marketing campaign.





-- 
Pure Personal Opinion | HoffmanLabs LLC