[seul-sci] For discussion

["Pete St. Onge" <pete@seul.org>]

First off, my appologies for silence on the list. I lost my net connection
for a bit right after the announcement, but this has been resolved.

I have a few throughts on this list and scientific uses of Linux already
drafted, and I have a will be posting these in the next day or so once
they're cleaned up a bit.

Perhaps the best suggestion I can make is for all of us on this list to send
a short intro message describing ourselves and out interest in Linux and
Science / Engineering and Research. I've already sent a message to the list
(http://www.seul.org/archives/seul/sci/Feb-2000/msg00003.html), but I'd like
to expand on it a bit.

I'm 28 years old, and have almost completed an MSc in Biology from McGill
University in Montreal, Quebec, Canada. My research interests are primarily
in physical limnology, and how lake habitat is affected by human activities.
I was fortunate to be involved with a fairly broad project looking at the
effect of forestry on aquatic systems. The downside (?) was the sheer volume
of data that we had to deal with.

For research users, our lab was exclusively made up of a couple of Win95
boxes used by some 30 people or so. Somebody had to maintain the machines,
and more often than not, it was me. I found the hard way that Win95 really
isn't a great research platform: no file permissions, no way to prevent
other users from installing software or deleting critical system files.
Taken together, this made for both a very unstable machine and my
corresponding improvement in the use of colourful lanaguge.

Just as important as stability, I found that site licences were getting
pretty darned expensive. Almost by definition, scientific users fall well
outside of the 'normal user' profile, so although we use spreadsheets and
word processors like 'normal users,' we're very often doing things that
conventional software suites cannot really support and thus the need for
specialized software (stats packages, graphing packages, etc). I noticed
with some packages (SAS, Sigma Plot) that their site licences could get
quite expensive - CAN$150 for a single year for one package! What I find
particularly unfortunate in some labs is the fact that potentially useful
research cannot be followed through on because the researchers do not have
access to the software required to carry out the task or the analysis.

Many of these packages have GPL'ed equivalents or, in other cases, a
superior replacement: For example, most of my research work using Linux thus
has centred around the R statistics package (http://www.r-project.org).

Linux is great for a lot of reasons: stability, price, applications. But
it's still far from perfect for my field - there are still a lot of gaps in
the available software that need to be filled, and I'd like to see these
gaps identified. Also, among my peers at least, there was a tremendous
resistance to using Linux in a lab setting as most of these people were
Windows users never having seen a Unix box before. These experiences struck
a chord with my Linux-using scientific colleagues and others as well, and
thus this mailing list.

That's pretty much it for me. Who's next?

-- 
Pete St. Onge
pete@seul.org