I've been revamping my website over the last few weeks. It's been
slow going, in part because there are a lot of behind the scenes
changes. The site is a set of files; there's no server except for
Apache. I could install a server, but the machine is hosted by Pair.com and running a server would
cost more, which wasn't warranted for a static site. When I updated
news items or released a software package, I made a copy of the
template and modified everything by hand. Tedious, but it was only
done every three months or so.
That's changed, as you can tell. I'm using the site more as an advertising medium, so people will know about my company and what it does. (If you want to hire us, contact firstname.lastname@example.org.) I replaced the hand-written templates with a templating system based on ZPT (more specifically, on SimpleTAL and a highly modified version of PubTal).
It took a while to get running, in part because I used it to experiment with new ideas, include ZPT. I also have a good idea now of how to write a Pythonic plug-in system, and tried out a few of the more obscure modules, like multifile (Hint: not worth it.)
Were I to do it again -- something I don't look forward to -- I would host Zope locally and write some code to export the relevant pages. Then again, someday I'll set up a server. That'll even let me have these "trackbacks" I read about. But since I read web pages and send emails, I still haven't drank that flavor-aide.
The next stage is to support RSS feeds. I looked around and didn't find any good code to generate RSS. The ones I found either didn't handle proper escaping, were incomplete (then again, most people don't need RSS), or just treated RSS as a generic XML and/or RDF. I wanted something which acted like Python objects. So I wrote PyRSS2Gen. But it probably won't be finished for another few days as I'm going to Michel Sanner's Computational Representations of Biomolecules workshop at UCSD followed by SciPy at CalTech.
I'm really looking forward to Michel's conference. It's something I've been interested in for many years. My first "computational representation' of a biomolecule was in 1992 when as an undergrad I helped write a parallelization of CHARMm for Paul Bash and since then I've been a major contributor to VMD, NAMD, PyDaylight and been involved with (either as contributor, user, or simple kibitzer) for frowns, OEChem, Tripos (they have some thoughts about using Python!), and others. Even more if you talk about software I've used which is based on molecules but where I haven't seend the API.
Andrew Dalke is an independent consultant focusing on software development for computational chemistry and biology. Need contract programming, help, or training? Contact me
Copyright © 2001-2010 Dalke Scientific Software, LLC.