[bip] welcome!

[Andrew Dalke]

On Jul 25, 2007, at 6:01 PM, Titus Brown wrote:
> I'm curious -- what does everyone do, and what do they use Python for?

Hi all,

   I'm Andrew Dalke.  I started using Python in 1998, but wanted
to use it a couple of years earlier.

   These days I mostly do chemical informatics, as a consultant,
for biotech and pharmas, and some academics.  My current biggest
project is an in-house molecular descriptor and model calculation
system.

   My background was in molecular modeling.  I then did some
bioinformatics, but as a consultant, chemistry pays more money.
I think it's because some time ago the chemists figured out
the alchemical trick of turning various white powders into money. :)

   The days I sit on the software implementation side of things.
My interest is in developing usable systems, where "usable" means
"human factors" not "can be be used if you get mad enough at it."
This can include API development and interaction design.

   Normally these go under the category of "architect", but I
implement, and break things, and do root cause analysis.  I'm
the annoying detail oriented person that says "on line 591 of
the format documentation you said XYZ but on line 25643 of
abc.dat it does this; probably because of a special case that
the output code didn't handle", and identifies buffer overflows,
and finds and works around software flaws so researchers won't
need to later on.

   I use Python because it's one of the few languages I know
of which are approachable by scientists (who typically have
relatively little programming experience) and enjoyable
by programmers (at least the kind aren't avid brace-aholics).

   For example, Tcl is easier for scientists to learn, but
most programmers get annoyed when building large systems in
Tcl.

   Oh, and I think I get to say that I'm a core Python member.
Outer core.  Very near the mantle.  But close enough that some
of my string optimization code is in the Python 2.5 release.
So if you're parsing GFF files you should see that code like

     for line in f:
         count += len(line.split())

is also about 15% faster. ;)



				Andrew
				dalke at dalkescientific.com