Chronology of a successful Y2k project (11 years to completion)greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread
From Westergaard's site.
Perhaps successful remediation can be done in less time by similar sized companies. Maybe even by larger ones. Then again, maybe not.
-- regular (email@example.com), April 28, 1999
I'm not a big fan of David Eddy's, because he usually takes 1 or 2 examples and tries to use them as blanket generalizations of the entire picture.
In this article, Eddy says that the Y2K effort, "..took 11 years, because that's what it takes to assure success of complex software projects of this magnitude." Is this truly the case? I would think that more than a few CIO's and CEO's would be a bit miffed over 11-year projects.
Second, regarding the Y2K work done, "Total staff hours through the end of 1998 was 20,300 (7300 in 1998 and 13,000 prior to 1998)." There are 55 employees in the computer department, and the computer development and management staff is under 30.
Using 1998, when the brunt of the work was done, assume that 15 of the 30 were used on Y2K work.
7300 hours/15 people= 487 hrs/person Assume 8 hour days (conservative) 487/8= 61 days. Assume 20 work days per month (This allows for 22 extra days per person for sickness, vacation, etc.)
Based on this, 15 people would spend roughly 3 months working on Y2K in 1998. This is 36% of the project. The previous 64% took 9 years to accomplish.
Just for fun, how long would it take 15 people working 8 hour days, 20 days per month, to do the entire 20,300 hours of the project?
20,300 total hours/15 people= 1353.33 hours per person 1353.33/8 hrs, per day = 169 days per person 169/20 work days per month = 8.5 months
All of the work is not completed at the same time. If we buffer this number with another 6 months, we are at 14.5 months. Is it possible that this project could be completed in a little over a year? Maybe less? The numbers seem to bear this out, and the numbers used here are all pretty conservative.
I'm not saying that I'm exactly right. But David Eddy's analysis seems a bit goofy. In this case, one company took 11 years to do a project at their convenience. Oddly enough, it took 9 years to do 64% of the project, but they were somehow able to do 36% of it in one year, and probably less, depending upon how many people were on the project.
I kind of compare it to back when I was in college. Some people might spend the entire 11 weeks of the quarter working on their final project. Some may spend 5, and some maybe only 1 or 2. But, in general, the quality of work and the hours spent are generally the same.
Different companies allocate different resources at different times to accomplish specific projects. David Eddy is probably right, it may have taken a company 11 years to do a Y2K project, but many companies are going to do it in a lot less.
Finally, note that the 11 years is an exaggeration. Their appears to be little done between 1988 and 1994.
-- CJS (CJS@CJS.com), April 28, 1999.
After reading the article, I had to wonder ... was this the same ASI that I used to work for? Probably not, since the company I worked for was much larger. However, I did notice the similarities between what the small company was doing and what we were doing.
The large company brought in a 4GL tool to redesign and reconstruct their systems (6 major application areas), over 5 million lines of code on one platform, in 1996. They are still working on it and have a programming staff of over 200 people.
Did they start late?? Yes. Will they make it?? Who knows. They keep redesigning areas, and can't seem to finish the project. Maybe if they had just addressed y2k instead of re-vamping the entire system, they might have finished by now.
By the way, that company supplies about 150 'Fortune 500' companies with software solutions.
Think about it.
-- Donna (firstname.lastname@example.org), April 28, 1999.