Ian, (01)
I strongly agree with that point: (02)
IB> My experience is that interviews don’t work, and you end up
> second guessing the subject. (03)
I also agree with the following point, but I would add some
qualifications: (04)
IB> If you can get your hands on legacy data (the dirtier the better)
> then I’d recommend analysing that rather than talking to users.... (05)
I would add that you should also look at *all* the legacy documentation
ranging from the initial reports, specifications, requirements, comments
in code, emails among the developers, manuals for end users -- including
everything going back 40 years or more. While you're reading those
documents, you should check for discrepancies between the implementation
and the documentation. (06)
Of course, that would take a bit of work. Therefore, you should have
the computers read the English documentation and compare it to the
actual implementation. (07)
> I wrote a (fairly light-hearted) paper for Cutter on this, see :
> http://www.cutter.com/offers/forensicIS.html. (08)
It's interesting that you wrote it for Cutter -- because my colleague,
Arun Majumdar, who implemented and used the tools to do the analysis
of *both* 40 years of legacy code and the English documentation, did
it while working on the project as a consultant for Cutter. (09)
For a brief summary, see slides 24 to 27 of the following talk: (010)
http://www.jfsowa.com/talks/pursue.pdf (011)
A major consulting firm estimated that the project of reading all
the documentation and relating it to all the legacy code would
require 40 people for 2 years = 80 person years. With the software,
Arun Majumdar and André LeClerc took 15 person weeks to do everything
that the consulting firm had promised -- and they also found the
discrepancies between the implementation and the documentation.
(They might have missed a few, but they found many that the human
programmers and users never found during all those years.) (012)
John (013)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (014)
|