[Pipet Devel] analysis management

J.W. Bizzaro bizzaro at bc.edu
Tue Jan 12 09:58:26 EST 1999

I thought I'd share some ideas I had about how the Loci user might manage
multiple analyses.  By management I mean keeping track of what has to be done
and what was already done.

The first idea I thought of some time ago but haven't mentioned yet.  It is an
expansion upon the concept of a log file.  Normally log files are generated one
for each run of a program.  But I think we can change that a bit to suit the
need of any scientist.

You know a "good" scientist will write all experiment data in a physical
journal.  I think though that it is most incovenient--a real headache infact--to
take everything that comes off of the screen and write it down.  Even cutting
out printouts and gluing them is can be a big hassle when you consider how much
data a computer normally generates. (I infact convinced my advisor to let me a
keep a computer-based journal--in HTML).

Well, to get to the point, I think Loci should keep a running log of all
actions.  That is, record everything to a single file--in HTML--with links to
data files, images, whatever.  Even keeping track of times down to the
second...better than anyone can do with a notebook.  Imagine having a
Web-browsable catalogged journal of all Loci analyses!

The second idea I mentioned already and have on the Loci Web page.  It has to do
with using icons and arrows to represent documents and analyses being performed
on them.  I've seen this before, although it is not very common.  What I wanted
to point out to you guys was the data mining program called "Clementine".  Has
anyone used it?  Ken Marx (Lowell professor that is entertaining the Loci
Project) told me the user interface works much like I am describing for Loci. 
So, here is the Web site for Clementine:


And here is the screenshot.  If nothing else, just glance at it to see what I am
talking about.


In other news, Prof. Marx says he will purchase a new Linux box to dedicate to
the Loci Project and act as a Web server.  We can each have accounts, etc.  I
also convinced him that when the Project takes off we may need more servers to
host the first server-side analysis loci.  Linux boxes are pretty cheap, so I'm
sure we'll get just about whatever we want ;-)

J.W. Bizzaro                  Phone: 617-552-3905
Boston College                mailto:bizzaro at bc.edu
Department of Chemistry       http://www.uml.edu/Dept/Chem/Bizzaro/

More information about the Pipet-Devel mailing list