Sent to massyr_at_washpost.com copied to classiccomp:
My cut on a nook in computer history. It's mine as I was only of
those nutty kids hacking a pdp-8 in highschool in 1969 wondering about
31 years lateer and how that would not fit into a 12 bit word used to
express date.
<i'm working on a piece here at the Washington Post about the Year
<2000 bug, and
<although there is plenty of info about what it is and how to combat
<it, there is very little
<history available.
<specifically, i'm trying to find out 1) who were the co-authors of
<the COBOL language
<along with Grace Murray Hopper,
Since she was military (LT at the time if memory serves) her fellows would
be military members as well. That may help you.
2) who decided to use two digits for the year instead of four,
Rather than answer that question directly as I cannot I'll supply resoning
instead. Computers over the years have come down in price, up in speed
and increased greatly in storage and memory. Those last two items are
significant to the discussion.
Starting in the early 50s a computer with 4000 words of memory and sorage
of say several hundred thousand words of data were large and scarce. The
technology for making memory and bulk storage were new. This would remian
true for many years, through the 60s. Because of the cost for even a few
bits of storage it was resonable to not store obvious data like the first
two digits of the date, after all tht was known and the lifetime of the
mache was like five years at best. So programming languages like Cobol
and more importantly the operating systems that managed the machine would
only keep the minimum amount of storage for the variable part of the date
(month, day, year). Memory was a premium for many machines and costly way
waste it on storing things that are assumed.
This habit persisted will into the '80s as while memories got larger and
machines smaller many would be expected to run the same or related
programs from their older parents. the scene is now set. Another thing
was the emergence of standard programming languages where certain aspect
of the language were expected to behave the same on sometimes radically
different machines.
The rest is history as many programming languages like cobol were expected
to be dead. However conservitive users like banking, government insisted
on tried and proven programming tools and languages thus propagating abd
prolonging the life of these burdend items. The burden was they had to
behave like their ancestors right down to little things like the date!
3) who discovered the Y2K problem?
It wasn't "discovered" nor was it a new problem. Some manchines and
operating systems had the equivelent of the Y2k problem several times
over and well before the year 2000. The example I know best is the
Digital Equipment Corp PDP-8, in teh mid to late 60s this machine was
introduced with the following attributes, small, low cost. It would
persist well in to the 1980's in various smaller and more compact forms.
The significance of the machine was it's the first "minicomputer" to be
sold in volume. In it's time thousands was big volumes. Now, it was
also a small machine in memory. The designeers for one of the popular
operating systems for it (known as OS/8) knowing that memory was a
precious commodity used the minimum needed to implment time and date.
The year portion was accorded a platry three bits meaning it could only
count from 0 through 7. Since this was introduced in roughly 1969 their
Y2k problem would occur every 7 years!. This was not the only machine or
company, just an example.
I know I was asking some of my friends what was going to happen as far
back as in the late 60' and early 70s when as high schools and college
students we were asking what happens if this machine should still be
running in the year 2000? Some of us considered that unlikely as the
pace for new machines at the time suggested it's life was maybe five
years and that pace was accelerating. Was was missed was the persistance
of some applications programming languages. While the PDP-8 has been out
of production for over 15 years many are still in use. While Cobol has
dropped as a mainstream language for over 10 years the problems that
are now tried and proven persist. This is true for several other
languages and some operating systems as well.
So no, the year 2000 problem wasn't discovered, save maybe for the media
person to finally put into print what every programmer and systems person
already knew. "We didn't use enough bits and were running out".
Allison
Received on Tue Apr 07 1998 - 08:48:54 BST