[Info-vax] How to Avoid Old Software, Old Bugs?

[Dave Froble]

On 11/13/2019 3:58 PM, Stephen Hoffman wrote:
> On 2019-11-13 20:11:47 +0000, Dave Froble said:
>
>> On 11/13/2019 12:49 PM, Scott Dorsey wrote:
>>> Stephen Hoffman  <seaohveh at hoffmanlabs.invalid> wrote:
>>>>
>>>> ...the usual and effectively-futile requests for re-debugging old
>>>> and long fixed bugs rather than staying current,
>>>
>>> There are many popular systems where one has a choice between old
>>> bugs that are fixed in newer releases, and new bugs that have been
>>> introduced in newer releases.
>>
>> There is the issue.  There is a lot to be said in favor of "if it
>> ain't broke, don't fix it".
>
> If we lived in the computing world where that was truly possible?
>
> Alas, we do not.

Well, some might, but I digress ....

> Not if there are network interconnections. Delegation and distributed
> authentication. Distributed logging.  Remote management.  Etc.

Not too many have stand alone applications, systems, and such any more.

>> As a developer, I'm running the latest from VSI.  The worst, and the
>> best, that can happen is I run into something and I report it to VSI.
>
> Which is the world we're all living with.  The world that we're all
> living in.

And, for disclosure, that is what we're telling CODIS customers to do. 
Support with VSI and the latest OS version.  If the sticky stuff hits 
the fan, well, that's what support is for.

Regardless, there is still some justification for "if it ain't broke, 
don't fix/break it".  That is just reality.  I personally do not feel it 
is sustainable.  But there is no "one size fits all", regardless of 
preferences and beliefs.

> Hopefully VSI makes few mistakes with their updates and upgrades.
>
> But then none of us can afford the cost of VSI making absolutely no
> mistakes, either.
>
>> However, a production system usually cannot afford to have problems.
>> One might attempt testing, but the real test of any such system is
>> actually using it.  So, it can be a bad decision, either way.
>
> The most predictable thing that'll happen to a single-server production
> system centered on uptime and that cannot have problems... will be
> problems.

Oh, my, yes.  Long ago, I learned that when someone tells me "that never 
happens", I should definitely allow for such a happening.

>> It's why someone was recently looking for a MiceoVAX 3100 Model 98.
>> They may have performed extensive testing and certification, and would
>> really rather not have to go through that procedure again if they can
>> avoid it.
>
> Various of us here want brute-force uptime.

It's not always just for uptime.  Consider the very expensive certification.

> App designs based on brute-force uptime will almost inevitably fail, and
> the apps and the systems will almost inevitably fall behind current
> software.

What?  Like VMS did?

> Various of us have configurations we don't even want to reboot.  What
> does that state about the fragility of our designs and of our apps?

Reboots don't bother me, and, while I can get careless on my development 
systems, on a customer system, a reboot always follows any changes to 
the OS.


-- 
David Froble                       Tel: 724-529-0450
Dave Froble Enterprises, Inc.      E-Mail: davef at tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA  15486