Re: Y2K

[Glenn Bacon <gcbacon@roadrunner.com>]

Michael C. Richardson wrote:

>>>>>> "Jay" == Jay Hanson <j@qmail.com> writes:
>    Jay> We migrated to IBM #1401.  It was a BCD machine -- not binary.  We
>    Jay> had to maintain compatibility with our databases -- card files --
>    Jay> thus, short dates.
>
>  Fine.
>
>    Jay> The cards for each year were stored separately, and we just changed
>    Jay> the program each year.
>
>    Jay> Habits die hard, the average life of a program back then was about
>    Jay> three months.
>
>  Like I said: "The systems were simply *NOT* engineered to last."
>
>  Had the system stayed on the 1401, there would not be a y2k problem.
>Why? Because "just changed the program each year."
>  The problem was when it went to a non-card system, it probably was
>not maintained anymore.

 Would you have engineered them to last if you were in charge?  Would
that have been a wise decision?  As I recall most 1401s didn't have
disks and their communications capabilities were at best latent.  Your
focus on "engineered to last" ignores the reality of the evolution of
computer architecture.  By the way, I don't think 1401s had a clock
that recognized years, do you remember?  I don't remember one on the
701s and 650s that I programmed.

It's a real stretch to blame the bulk of Y2K on gain-max.  (Unless, of
course, your ideology precludes other analyses.)

Glenn Bacon