It's expensive for businesses to write good programs. If
they aren't going to sell many copies (MacVector is an
obvious example) it needs to be very expensive.
Supporting people who write freeware programs is a great
way to solve the problem. In my gene-linkage work, I use
linkage, fastlink, simlink, GCC, peddraw (the mac one),
paradox, endnote, mirosoft fortran, etc...
The first five of those are available free or at nominal
prices ($10-$35 for peddraw--that doesn't pay any salaries).
There are many programs I would like to see in freeware.
Archie (the server code, not the client code) would be
great. Ever wonder why there aren't many archie servers
(for indexing files at anonymous ftp sites) out there?
The reason is that the archie-server code costs $6,000
(although there is a substantial discount for educational
institutions).
I think that many programs that we'd like to see (including
a freeware version of archie) would be the perfect thing
for a computer science software engineering class to write
as a team project.
Remember that the original version of spice (the #1 electronic
circuit modeling program) was written by a class of EE grad
students twenty years ago. They were told "if you write the
best electronic circuit simulator program ever, you all get
A's."
They all got A's, I think the prof eventually got tenure, and
the computer industry made hundreds of millions of dollars
building computers to run spice.
If you're a professor, and would like to see a program written
to do a particular task, write down a complete specification
(with examples, suggested algorithms, etc...) and drop by
your local CS department. This will also reinforce the
importance of CS professors (in the minds of the tenure
approval committe?) who choose to write (or sponsor student
projects to write) software useful to us.
Of course, now that I've said that I can feel guilty for
not having had any time to work on emaillink. Ugh.
Darrell Root
rootd at ohsu.edu
