Re: Request for reuse tool info

[saimiri.primate.wisc.edu!zaphod.mps.ohio-state.edu!wupost!cs.uiuc.edu!joh nson@ames.arc.nasa.gov (Ralph Johnson)]

In my opinion, the need for library tools is grossly overstated.  The
hard part is not the library tools, it is the library.  The fact that
everybody wants library tools is a symptom of the computer science
disease of trying to cure every problem with a new tool.

I agree with Greg Aharonian that you can probably just use any old
searching tool, like a relational database or a text retrieval tool,
if you had a decent database to search.  Greg's got a database, so
why doesn't someone pay him to put it into a form that can be searched
properly?  Of course, if someone does, they'll probably find out that
the database is missing some important information.  So, pay Greg to
put that information in.  It should only take an iteration or two to
get it right, it will take a lot less effort than writing some big
program, and it is likely to work, while I don't believe that a new
tool will solve any real problems.

The database is a much bigger problem than the tool.  New software is
continually coming out, and someone needs to try it and find out what
it does.  There should be testing labs, people worrying about how to
classify software, etc.  It is obvious (to me, at least) that it is
better to have a single (or a small number of) companies do this and
sell the results than to have everybody duplicate it on their own.  
Somebody like Greg should sell rights to his database, and should use
that income to improve it.

I don't know Greg, though I have seen his name on papers and know a
little about what he is doing.  If I've got him figured out right then
he is trying to get into this business.  If every company that is 
interested in reuse would just sign a contract with him to provide a
database then I bet we would hear a lot less complaining about not being
able to find reusable software.  Once there was a good database, I bet
it would be a lot easier to make tools, too. 

Of course, organizing reusable software is not easy, especially when
you think of it as just a bunch of parts.  That is the wrong way to
think about it, naturally.  The right way to think about it is as
frameworks.  If you think I wrong, please get my paper by anonymous
ftp from st.cs.uiuc.edu in /pub/papers/reusable-oo-design.ps and tell
me what is wrong with it.  (Or use send e-mail to the archive server
archive-server@st.cs.uiuc.edu and say "send papers/reusable-oo-design.ps").

Ralph Johnson -- University of Illinois at Urbana-Champaign