tom wrote in
. net:

Incorrect.

The Y2K bug demonstrated that in many cases companies wouldn't listen to
engineers, or more correctly, programmers. Some did, most didn't until
very late. I was working on Y2K upgrades in 1991 at utility companies.
Many other, less critical types waited until much later.

Most programs that really counted, like utility companies billing and
control programs, were written in the 1960s and no one ever expected
them to be used for even 10 years without replacement, let alone 30+.
To blame engineers is foolish, since they had little to do with the
programs in the first place, or the programmers, who could not foresee
the future.


This is true but it doesn't render an 'opposing' view false. People do
various things, some change with fashion, maybe don't beleive they're getting
value unless it's changed whimsically, bloated to look bigger than the
competitor's offering... Others (like me) prefer small efficient modular
systems whose parts serve long after and beyond any intent because their
makers managed to focus on a singular issue while not losing sight of context
while they worked, in effect future-proofing their design to some extent.
This lets me adapt things where they won't just continue to work as is.

I think the Y2K 'bug' wasn't corruption, but folly. It only takes a few more
bits to store the full year values, yet they were not used. No doubt some
engineers did think of it. Similar problems happened in hard disk addressing.
I think people are quickly learning that allocating storage for large
addresses is MUCH cheaper than fixing it too late, several times a decade.
Network addressing with IPv6 for example, that one might never have to change
in all our future. We could populate the local group of stars with people
before we ran out of addresses that way. (Although I guess many people are
more likely to want to network toilets and toasters instead). Even Microsoft
deserve a bit of credit with their 2-byte text allocations filled for years
with alternating ASCII and zero bytes. Looked stupidly wasteful to me all
those years but it's paying off now that Unicode usage extends so much. They
evidently learned from Gates' assertion that 640 KB of RAM was more than
enough for anyone.

While it's not easy to forsee the future, it IS easy to forsee that modest
size addressing systems will be outdated fast, because any successful system
gets adapted to more uses even if the total number of users doesn't change.
MIDI, used in music, has become a versatile machine control language, mainly
because its inventors did think of this, in 1983, and MIDI continuously
outlives its huge number of obituaries.