Here's my two cents on this thread.
1. If you're setting up any kind of sequence analysis center keep in
mind that the hardware costs are usually negligible compared to the labor
costs. The only time when this is not true is when system management
is by slave labor (ie, grad students). Personally, I think that
the various granting agencies, and certainly the universities should
outright forbid that practice, but I'm not holding my breath.
2. Software costs can be pretty misleading. "Single seat"
software (like on a PC or Mac) is usually very expensive when you
do any sort of cost/user analysis. For this reason, GCG is a great
bargain at 3000/year (especially with > 100 users) compared to the PC
and Mac options. This also shows up in service contracts and the like
- it's cheaper to maintain a couple of Vaxes or Unix boxes than an army
"Free" software is not without costs - maintenance on this stuff can be
quite a pain or a breeze, depending entirely on the whims of the
program's author. Sometimes you can get bugs fixed and documentation,
other times, no.
3. One system manager usually costs less than N "part time" computer
honchos distributed at one/lab. However, there is a strong tendency to
budget for the latter and not the former. There is an even stronger
tendency (apparently among granting agencies) to approve money for
hardware but not for labor. On this last point I'd like some feedback:
it's been reported to me that this gets lined out on grants, but I've
no personal experience with it.
mathog at seqvax.caltech.edu
manager, sequence analysis facility, biology division, Caltech