home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.lang.c:16786 comp.software-eng:4398
- Path: sparky!uunet!spool.mu.edu!sol.ctr.columbia.edu!destroyer!ncar!noao!arizona!optima.UUCP
- From: rws@optima.UUCP (Ronald W. Schmidt)
- Newsgroups: comp.lang.c,comp.software-eng
- Subject: maintenance questions
- Message-ID: <26750@optima.cs.arizona.edu>
- Date: 19 Nov 92 09:21:15 GMT
- References: <1992Nov19.052343.14116@netcom.com>
- Sender: news@cs.arizona.edu
- Followup-To: comp.lang.c
- Lines: 46
-
-
- My initial posting of some questions got a few answers, some
- agreeing, some disagreeing - both helpful. What I was really trying
- to assess was the potential utility of a MYCIN type symptom based
- expert diagnostic system for software maintenance. If maintenance is
- carried out by the original authors of the system then a symptom based
- diagnostic approach will not be very helpful. Human intelligence will
- be better than expert intelligence. The utility of such an approach
- is clearly most useful in a situation where people whose familiarity
- with the system is not so great must do maintenance and where the most
- frequent resolution to a problem is to match symptoms against known
- problems and give known solutions. I worked under such circumstances
- and am wondering how common they are in the software industry as a
- whole. I would expect that such circumstances would be most likely
- at places that produced software which had a wide distribution so
- that there is no chance for the original developers to handle all
- questions.
-
- I did manage to find a book with more recent maintenance statistics
- than Boehm et al. Capers Jones book on _Programming Productivity_
- gives some helpful statistics. Unfortunately the impact of modern
- programming languages on field maintenance was only touched on
- and most of the answers to my questions were not directly addressed.
-
- For any who did not care to comment the first time, or would
- like to try again in the above mentioned context.
-
- 1) In a competitive market there is a need for short term corrective
- maintenance that is not quite like software evolution. There
- is a need for fixes that get a customer around a problem safely
- and immediately to prevent the customer from switching to another
- product.
-
- 2) Companys hire special group of people for maintenance.
-
- 3) In corrective maintenance creating new fixes is burdensome and undesireable
- and infrequent compared to the distribution of already known
- inter-release fixes.
-
- 4) By the time the decision is made to modify software a lot of information
- has been gathered by a support person attempting to avoid making
- any modification.
-
-
- Thanks in Advance
- -Ron
-