On Wed, 17 May 2000, Ralf Baechle wrote:
> This world also has an economic part and there a machine that is written
> of in the books, that is after just a few years, has no more value and
> even becomes a burden. So in most companies perfectly usable and even
> adequate configured machines get replaced. At this point the manufacturers'
> interested in supplying services also decays because there is just no more
> sufficient customer demand and economic justification to keep spare parts
> (read: dead capital) and trained personal. For private customers it's
> not that much different, usually the machine goes to the dustbin when
> the newest slash & kill 3D games are turning into a slide show. So
> after just a few years most manufacturers will flush everything including
> the old documentation down the drain. Add the fact that the companies
> of interest are mostly US located where people change their job more
> often than their pants ...
Of course -- but I do think that the cost of retaining docs is minimal,
at least when compared to profits gained from the respective equipment.
Upon the commercial end of life of some stuff an action of preserving docs
might be performed by the team that used to support it. I don't think
putting docs somewhere on an FTP server takes more than a few days and
then hard drives are so cheap there is no point in not keeping such data
> Specs are a bit special, though. You wouldn't believe how many high
> tech specs don't exist in a form that is publishable. That's very common
Actually I do understand it -- I've never been much enough inclined to
work on docs as opposed to making real stuff. As long as they are not
handwritten notes on unsorted pieces of paper, but some sort of electronic
document, which I would expect to be the case these days, they are still
> in places where the authors of the specs and it's readers are close to
> each other, that is in traditional system companies such as Digital, SGI,
> Sun or HP. Their motivation to actually produce high quality documentation
> used to be relativly low since at the time of product development the
> assumption was that they'd never ever publish the specs.
I do not need a high quality docs -- I think anything that lets deduce
component's operation is much enough to make it useful. And these bits
must be written somewhere anyway, as I wouldn't treat human's mind perfect
enough to be the only source of technical details. It's much too
> Free operating systems are changing the rules of the game, so expect
> the way documentation is handled to change as well.
Well, better late than never, but there is still plenty of excellent
equipment out there which loses against crappy old PCs because of lack of
+ Maciej W. Rozycki, Technical University of Gdansk, Poland +
+ e-mail: firstname.lastname@example.org, PGP key available +