drw@culdev1.UUCP (12/07/87)
At present I'm trying to get a manager to look at improved software development methods. She believes that what are considered "improved" methods just increase the time and money required to finish the project. (She previously worked at a government contractor; this may have something to do with her attitudes.) I need to find references to both (1) software companies (preferably in the Boston area) that do things right, and (2) real examples of projects that saved time and money doing things right. Please e-mail replies. Any help would be greatly appreciated. Dale -- Dale Worley Cullinet Software ARPA: culdev1!drw@eddie.mit.edu UUCP: ...!seismo!harvard!mit-eddie!culdev1!drw Nothing shocks me -- I'm a scientist.
rion@wdl1.UUCP (Rion Cassidy) (12/11/87)
Here in the bay area we have a mass trasit system called BART (Bay
Area Rapid Transit), probably you've heard of it. Yesterday I heard
on the news that a new computer control system for BART was five years
behind schedule and $25 million over budget. While I have no idea of
who is doing this work and what methods they have employed, I think
that this should serve as an example that there are large software
projects currently under progress out there that need help, and
software engineering (of some sort) could easily be that help.
A couple of months ago I went a seminar put on by a local ACM chapter.
The speaker was Tom Gilb, and from the course material I have found
what I believe some of the more significant references.
M. E. Fagan, "Design and Code Inspection to reduce errors in program
development" IBM systems Journal, 15, No. 3, 1976.
Mills,Hardin,Dyer,Michael,Quinnan: Articles on Evolutionary Delivery,
IBM Systems Journal, No. 4, 1980.
Wong Carolyn, "A Successful Software Development", IEEE Transactions
of Software Engineering, Nov. 1984.
Gilb, Tom,"Software Engineering using Design by Objectives", ACM
Software Engineering Notes, April 1984, pp.104-113.
Rion Cassidy
rion@ford-wdl1.arpa
...{sgi,sun,ucbvax}!wdl1!rionfutor@lll-lcc.aRpA (Randy J. Futor) (12/17/87)
> rion@wdl1.UUCP (Rion Cassidy) writes: > . . . Yesterday I heard >on the news that a new computer control system for BART was five years >behind schedule and $25 million over budget. While I have no idea of >who is doing this work and what methods they have employed, I think >that this should serve as an example that there are large software >projects currently under progress out there that need help, and >software engineering (of some sort) could easily be that help. You should look into the writings & teachings of Steve Mellor, now of Project Technology, Inc. in Berkeley. He worked on the BART software that's now running the trains (maybe it was a version ago instead...) & he has some interesting stories to tell. He was also my officemate while consulting on the EDS project out here in laserland so I heard some of them more than once. One of the biggest problems he's had (as well as any number of us that embrace some of the new methodologies) is getting the higher-ups that think 'hack & iterate' is a resonable way to write a large piece of software to listen when you give them a time&$$ estimate of their proj- ects. It's difficult to convince people who've seen code written that it's not real clever to try to sit down & code a large piece of soft- ware anymore & that methods exist which allow one to PREdocument a SYSTEM (emphasis mine) & that these methods will drastically cut down system vs. documentation disparity as well as they cut down on main- tenance of the resultant code. It's a very large problem area & I could go on for more screenfuls than either of us would want to read; I suggest you contact PTI (or their equivalent if you prefer another method) if you're really int- erested in this sort of thing. If you decide to read through the 3 volumes of Ward & Mellor first, bear in mind that the methods have evolved & improved since their publication. -- Randy futor@lll-lcc.llnl.gov