[oclug] How open nature of Linux can work against it.

[Robert Brockway]

On Sat, 12 Aug 2006, Bill Strosberg wrote:

> Validation is just window dressing and ass covering for PHB executives,
> it has no application in the real world.  It give executives someone to

Oh definitely.  Unfortunately it is still present and a requirement for 
some purchasing decisions.

IMHO validation presumes a level of maturity in the IT industry as a whole 
which is not yet present.  Our industry is all of 60 years old - I expect 
it won't start to approach true maturity until long after I have retired.

> Hardware validation is worse, as you can not purchase the exact same
> system five minutes apart in the real world.  Parts change.  You can't

I was thinking about this again recently.  I'm sick of hardware being 
pushed to market so fast that the firmware is often buggy.  Debugging a 
firmware problem can be very timeconsuming (if it is feasible at all).
I strongly dislike end users being treated as beta testers for new h/w.

I'd really prefer h/w vendors to slow down a bit and start providing 
better tested equipment to the market.  Unfortunately this seems to be 
infeasible in the current market.  I simply avoid buying bleeding edge h/w 
and allow other people to do the beta testing.   I'll pick up the 
motherboard when it is on its 3rd or 4th bios revision (and is cheaper 
to boot :)

> design a board today that can be built two years from now.  Parts people

Yes this is very annoying from the POV of wanting to maintain consistency 
across systems purchased at seperate times.  It is almost impossible.

This comes back to the maturity I mentioned: If the industry as a whole 
was serious about producing stable computing environments we'd take a 
breath, put the brakes on, and focus on producing more robust rather than 
faster systems every other week.

> The problem with a static testing criteria is that real world problems
> never seem to use the testing methodology.  In  the real world, problems
> crop up from issues outside the validation criteria, again and again.

In that case the testing criteria are inadequate and need to be fixed.

This takes time of course - time that can't be dedicated because the 
engineers need to get the next motherboard to market :)

> If I owned a make of car that ONE GUY in the world could legally fix, or
> I could choose another brand that thousands could legally fix, I would
> choose the latter, not the former!

Indeed.

A great historial example of open standards is the use if screw threads. 
Up until the 19th century screw/screw driver designs were unstandardised.

Each machine shop used its own standards and the only feasible way to get 
new tools for a particular machine was to go back to that machine shop 
(vendor) to purchase the tools (that were no doubt offered at an inflated 
rate).  Imagine trying to surprive today without standardised screws and 
screw drivers.

Rob

-- 
Robert Brockway B.Sc.        Phone:          +1-905-821-2327
Senior Technical Consultant  Urgent Support: +1-416-669-3073
OpenTrend Solutions Ltd      Email:          support at opentrend.net
                              Web:            www.opentrend.net

If you are emailing regarding an open ticket please consider
mentioning the ticket ID as this will assist us in responding
as quickly as possible.