From: http://www.computerworld.com/home/news.nsf/all/9808123y2k a Cap Gemini survey...

Excerpt:

"Perhaps most startling is that half of the companies and agencies surveyed don't plan to test new software code and systems to make certain that they are truly ready for the year 2000. "That's a disturbing conclusion," Woodward said, noting that testing is crucial to make certain that year 2000 issues have been fixed and are not simply "in remission."

I think this is utterly expected. Some systems can't be tested, i.e. the phone systems, and others are simply not going to have time. As a developer I can tell you something --- NOBODY puts a system out without testing unless they are utterly insane. Not only do you not know how it's going to act --- but you have no idea of how it will impact your other, existing systems! This is sheer lunacy - and I wonder how they are planning on defending this deviation from standard software methodology to those folks coming with attorneys in tow seeking damages from them?

This will be really indefensible.

-- Chana Campos (chana@campos.org), August 12, 1998

Answers

I suspect that not all will agree that testing is the rule. In fact some might conclude it is the exception. With 20 something years of analysis/programming as background, I would suggest that there is a lot of 'fly by the seat of your pants' programming going on. Furthermore I will observe that most developers (at least in the business community)do not have engineering backgrounds. When they do actually test, the test plan (test plan, what is that?) is woefully inadequate.

(Besides will management delay an already late project for something as non productive as testing?) As for Y2K, corrections without testing MAY be better than no corrections at all. Only time will tell.

Am I off base in my observation? See the Information Week column titled Satisfaction Not Guaranteed at http://www.techweb.com/se/directlink.cgi?IWK19970929S0086

May your programs always be tested.

jh

-- John Hebert (jhebert@co.waukesha.wi.us), August 13, 1998.

I agree wholeheartedly that, for the most part, lip-service only is given to serious testing of software...heck, look at some of the stuff that's put out by the big boys! On the other hand, in a litigation situation - I'd be willing to bet that attorneys are familiarizing themselves with "best practices" and "quality assurance standards" even as we speak. While "usual and general practice" will be something of a defense - the basic mechanism of software development methodologies has been around for some time. And they're going to wonder why they ignored it.

I also agree that (hopefully) some repair is better than nothing. I also agree that there is probably (come to think of it - I've seen it..!) fly-by-the-seat-of-their-pants programming going on! I just wish I could forget about that statistic that says for every bug you fix another shows up somewhere else!

-- Chana Campos (chana@campos.org), August 13, 1998.