On Fri, 5 Dec 2003, Derek M Jones wrote:
> > I think most software engineering academics are acutely
> >aware that they are criticized in industry for doing irrelevant
> >research.
> I would not characterize it that way. I think the idea that academics
> do research that one day may or may not be useful in 10+ years is
> generally accepted.
I think we must go to different SE conferences or get
different messages from them.
> It is the claims that past and current research (at
> least in software engineering) is relevant today that causes much disbelief.
I do hope that that wasn't the sweeping generalization
about all SE research that it appears to be!
> > Many try to make contact and find it very difficult.
> Yes it is hard. But researchers in other domains have equal
> difficulties, but don't cope out (well perhaps some do, I don't know).
What domains are these, and how do they avoid copping out?
For instance, anthropologists can do a lot of good research in
companies by just observing and asking questions. Slows the
workers down a bit but is mostly unobtrusive. Experimental
psychologists can do a lot of good research using undergrads as
subjects, because they may be better able to make the assumption
that the undergrads are representative humans for the purposes
of what they are studying. SE researchers often have to go to
companies and get people to do things, and therein lies the problem.
> I think software engineering academics (come to that, all people
> investigating this issue) are trying to run before they can crawl.
> We need more basic research before studies involving so many
> variables can be carried out.
I suppose an ideal study all round would be one that is
well controlled, that takes a very small amount of time per
subject (say less than 90 minutes, so that a developer can do it
as a diversion and not worry about what line item to charge it
to), that takes a very small total investment of time by the
company (say 2 person-days, so that it's almost like someone
calling in sick), and that yields publishable results.
Having to work within these constraints seems to mean
basically avoiding certain areas of research. For instance, any
tool or technique that takes more than 60 minutes to learn is
basically out -- even though companies routinely send people on
1-week training courses for proven technologies. How to vault
the gap between "worth 1 hour" and "worth 40 hours" is a huge
puzzle for researchers, especially if they have to also extract
a continuous stream of publications out of the research.
> > If so, please put me in contact with this person.
>
> You must be joking. I would keep them a very closely guarded
> secret for my own experiments.
:-)
> From the companies perspective there does not appear to be much
> return on investment for them. I think the secret is to appeal to
> developers direct, to take part in their spare time.
I agree on that point. I would just say that it does
impose a large number of constraints on the research area under
study.
--Jamie Andrews.
Associate Professor
Dept. of Computer Science
Univ. of Western Ontario
London, Ont. CANADA N6A 5B7
[EMAIL PROTECTED]
http://www.csd.uwo.ca/faculty/andrews/
[p.s. anecdote about subject line: I first arrived in
Edinburgh to attend the university in late September 1987.
Sometime around Oct. 25 I got a knock on the door of my flat.
I opened it to find three local kids in rudimentary costumes
with soot on their faces, holding bags.
KIDS: Money for the guisers.
ME: Pardon?
KIDS: You from America?
ME: Er... Canada actually.
KIDS: Ye have Halloween there?
ME: Yes. (?)
KIDS: We have something like that here. Guy Fawkes Day.
ME: Er... I know... but... isn't that Nov. 5?
KIDS: Aye we're early.
ME: Um... OK. (runs and gets some change)
]
----------------------------------------------------------------------
PPIG Discuss List ([EMAIL PROTECTED])
Discuss admin: http://limitlessmail.net/mailman/listinfo/discuss
Announce admin: http://limitlessmail.net/mailman/listinfo/announce
PPIG Discuss archive: http://www.mail-archive.com/discuss%40ppig.org/
