Re: Boeing and Dreamliner

[aek@vib.usr.pu.ru (Alexander Kopilovitch)]

Hyman Rosen wrote:

>The Araine 4 programmers exactly reproduced the Y2K problem in microcosm.

Why *micro*cosm? All Ariane rockets are for normal-scale cosmos, I think -:)

>They wrote code that took advantage of limited input range, and sent it out
>into the world with insufficient protection against the future. It was
>perfectly good in the same way as two digit years.

With these words you stepped on my (programmer's) native soil -;) , and here
I simply *know* why those 2-digit year representation was right thing in many
cases that old time (so programs that use it may be called perfect).

It's quite easy: imagine 1970th, IBM/360 running DOS/360, 64K or 128K of
memory, 2 or 3 hard drives, 7.25 Mb each, and 3-4 tape drives (about 20Mb
each). Then imagine that you must process all documents for all trucks in
quite big (5M) city every day. Each document contains several dates.

I hope it is obvious enough that several percents of whole data volume was
significant weight, and that there weren't many people who prefer to carry it
(having very limited hardware resources) for the reason that those programs
may survive 20+ years, and then, in far future there may be problems with
this issue.

We were all sure that those our programs can't survive 20+ years.
No one can predict in 1970th and early 1980th that hardware will skyrocket
while software will essentially stall. So, if there was an error, it was in
the prediction (or expectations) of general trends in computer development,
not in design of particular programs.

(Incidentally, here in Russia we did not feel any significant impact of Y2K
-- just because our mainframes did not survive that far; at the same time
many programmers here in 1999-2000 were making their living from American and
European Y2K problems -;)



Alexander Kopilovitch                      aek@vib.usr.pu.ru
Saint-Petersburg
Russia