IP: A bit more on Y2K and the past

[Dave Farber <farber () cis upenn edu>]

I cannot agree more with Bobs comment re focusing on the deisgn!!. CFG (Caine, Farber and Gordon)( http://www.cfg.com) 
for years marketed a product called PDL which focused at improving the design part of programming rather than the 
coding part on the grounds that that was where the real costs were. It was very successful in very large products where 
the lack of tools that allowed people to work with people and customers beccame more obvious and critcial.

Dave

From: "Robert Fabian" <rfabian () rfabian com>
Organization: Robert Fabian Associates
To: Dave Farber <farber () cis upenn edu>
Date: Sat, 14 Nov 1998 11:52:51 -0500


Dave

Your personal comments about Y2K prompt a few personal comments from 
me. I remember back to the thinking machines of the 1950s (Univac I, IBM 
650), and was actively involved with several installations in the 1960s.

I find most interesting a comparison between the cost of storage back then 
and today. By my rough calculations, there is a cost difference of 
(approximately) 1,000,000 between the cost of memory in the 1970s and 
today. Back then, core memory was priced at something like $1 per byte. 
Today, semiconductor memory is about $1 per megabyte. The same cost 
comparison holds between early DASD and current disks. 

Would it be right to use an extra few megabytes per record, today, to avoid 
a problem that will occur well into the next millennium? My strong sense is 
that today's designers would reject this kind of "unnecessary" overhead.

The lesson that I extract from our Y2K problems is not that we should 
consider all future possibilities. Remember, back then, we "knew" that the 
code was going to be replaced within seven plus or minus two years. It's 
never possible to take into account all possibilities. What we should do, 
however, is work toward designs that are robust enough to accommodate 
change, including change that has not been pre-planned.

The important lesson is about design. We're learning, but it's painfully slow.

Bob Fabian

On 14 Nov 98, at 11:19,  Dave Farber wrote:
<snip>
> I cannot ever remember worrying about the year 2000. It was out of mind as
> we struggled to get programs running and tried to keep up with the
> evolution of OS and language changes that flew at us.
>
> Sorry people
>
> Dave
> _____________________________________________________________________
> David Farber         The Alfred Fitler Moore Professor of
> Telecommunication Systems University of Pennsylvania Home Page:
> http://www.cis.upenn.edu/~farber     



---
Robert Fabian Associates  416-769-1885 71 Evelyn Ave Toronto M6P 2Z2
                     http://www.rfabian.com