I just can't hold back any longer. Although I'm not a systems administrator, I
do have a CS degree, and I have picked up some biology in more than five years
of working closely with MDs and molbio researchers. For the last three years,
I have done machine learning work strictly in molecular biology domains.
Some good points have been made, but some important issues have been left out.
Important issue number 1: There is plenty of need for people with both rigorous
CS backgrounds and some biology training, and for people with rigorous biology
backgrounds and some CS training. And there are some good new programs coming
on line that offer the resources to provide a good education in both fields.
Important issue number 2: There are many different CS skills, just as there are
many different lab skills. People who are good programmers may not be good
algorithm developers, who in turn may not be good systems administrators, who
themselves may not be good at training other users or selecting existing
software, etc. etc. etc. And different computer science degree programs train
people differently. I share Rick Westerman's consternation at the dominance of
computational complexity theorists at some computer science departments, and am
the first to admit that many PhD computer scientists never write another
program in their lives. On the other hand, people with degrees in CS usually
learn something about programming during their education... Although you
should remember that programming is not the most important part of making
computer systems work for biology. Access to programs (data sources, etc.)
that already exist, and the ability to exploit them is what most biologists
need, not the ability to write their own code. Some biologically sophisticated
people do have to want to write code (and there should be both CS and bio
career paths for us/them), but you don't have to be a hotshot programmer to put
together a very useful biocomputing system. You do have to be a damn good
system admin to handle a lab with a VAX, 2 Suns, 10 Macs and a dozen PCs,
handling a zillion different database formats, updates every quarter, lots of
PD software, and a bunch of users who don't have time to be trained adequately.
Important point number 3: It is extremely difficult to successfully write and
maintain a complex computer program without some serious training in computer
science. Hell, it's not that easy if you have a bunch of training and have
real experience in writing large systems. I have delved into the heart of a
lot of code written by biologists who have "picked up some computer science
along the way," including some very well known ones. The code usually works,
but there tend to be certain kinds of problems that make these programs hard to
extend, reuse or maintain. I'm the first to admit there are plenty of holes in
my semi-self taught biology. It's a good idea to be realistic about the
abilities of semi-self-taught programmers. Learn the lessons of physicists,
who are often saddled with keeping ancient computer systems running, because
they need some piece of software that only runs on that machine and no one can
possibly port it without rewriting it from scratch.
OK, now that I've gotten that off my chest, I can go back to work with a clear
conscience.
Larry
--
Lawrence Hunter, PhD.
National Library of Medicine
Bldg. 38A, MS-54
Bethesda. MD 20894
(301) 496-9300
(301) 496-0673 (fax)
hunter at nlm.nih.gov (internet)
