3.331 software development (87)
Willard McCarty (MCCARTY@VM.EPAS.UTORONTO.CA)
Sat, 5 Aug 89 15:30:21 EDT
Humanist Discussion Group, Vol. 3, No. 331. Saturday, 5 Aug 1989.
Date: Thu, 3 Aug 89 23:04:32 EDT
From: sdm@cs.brown.edu
Subject: Developing Software
Charles Ess pointed out that "it only took some five years and a few
million dollars to develop Intermedia for the Macintosh." This is probably
true, although I feel compelled to point out that at least some of that
time and money was invested in trying to make an IBM PC/RT look like a
Macintosh, and of course some of it was spent producing reports for the
funding agencies. Nonetheless, the undertaking was a substantial one, and
the end result is an industrial-strength hypermedia system.
Still, it's interesting to ponder how much could have been done if the
project were not so ambitious. For example, imagine that you wanted to
develop a primitive hypermedia system on a Unix workstation that ran the X
window system, and you had access to an experienced programmer who was
familiar with writing programs to run under X. Imagine further that you
were willing to settle for relatively small amounts of data in your system,
and you were willing to limit yourself to textual entries and bitmaps, plus
fairly simple linking schemes. In other words, you just wanted to "play
around" with a simple hypertext system.
My guess is that you could have such a "toy" system put together and
available for experimentation within no more than a month (exclusive of the
entering of the data for the system, which is a separate problem). What
would you give up in comparison to a system like Intermedia?
1. Ease of use. The user interface wouldn't look like a Macintosh, it
would look like whatever X happened to offer. This would make it
less suitable for inexperienced computer users.
2. Robustness. There would be more bugs in the system.
3. Views of the data. Intermedia offers web views and timelines (and
possibly others), and you wouldn't have those kinds of views.
4. Extensibility. Intermedia can handle large quantities of data.
Your system would become slower and slower as you added more data to
it.
5. Documentation. You'd have to learn how to use the system from the
programmer.
My point here is not that everybody should be writing their own hypermedia
system, but that the essence of such a system can be implemented in a
fairly straightforward fashion. What costs you is getting the system into
shape for use by outsiders. If you're willing to live with something that
is clearly home-grown, you can develop things fairly quickly and
inexpensively.
My second point is that many colleges have an ample source of programmers
who could potentially write systems such as I've described: undergraduates,
especially majors in Computer Science or a similar field. Certainly the
hypermedia system I described above could be implemented by an
undergraduate in about a month, presupposing s/he had the qualifications
I've listed above. Here at Brown University, for example, the CHUG
(Computing in the Humanities User's Group) and the CSDUG (Computer Science
Departmental Undergraduate Group) are exploring the possibility of having
CSDUG members work on implementing projects that are suggested by the CHUG.
The idea is that the CSDUG members gain experience with "real" computer
programs and with "real" research projects, and CHUG members gain
(potentially) useful software for the cost of advising a CSDUG member. As
far as I know, this cooperative effort is still quite tentative, but the
idea, I believe, is a sound one. If other Humanists have experience with
this kind of collaboration, I would appreciate knowing about it.
My overall point is that it need not be that difficult to develop software
for use in the humanities, whether on new machines or old. At many
schools, there is even a pretty good chance that you could coax someone
into writing your program for you. There being no free lunch, however,
you've got to give up something, and what you'll probably have to give up
is the luster of a polished professional-quality program. For some
Humanists, I suspect, the tradeoff may well be worth it.
Scott Meyers
sdm@cs.brown.edu