Eh, some of the items in this list aren't so great.<p>simplejson is included in the standard library as json now. Yes, cjson is faster, but it's also more fragile and, well, less standardized. Everything but simplejson is effectively dead: simplejson won the json library wars.<p>html5lib is an odd choice for a new HTML parsing kit. lxml.html offers its own fast and good HTML parser, and can also use html5lib or BeautifulSoup as its backend while still producing an Element tree with plenty of convenience methods (like xpath). I am not sure why html5lib would be better here (maybe it's not as complicated, somehow?)<p>As for MySQLdb, that thing is awful. There are some replacements out there, and out of them my personal preference is for oursql. The creator/maintainer is a regular (an op, actually) on the #python IRC channel, so it's easy to get support, and it fixes a lot of the nasty quirks of MySQLdb.<p>His first four suggested libraries are definitely good, though<p><a href="http://docs.python.org/library/json.html" rel="nofollow">http://docs.python.org/library/json.html</a>
<a href="http://simplejson.googlecode.com/svn/tags/simplejson-2.1.0/docs/index.html" rel="nofollow">http://simplejson.googlecode.com/svn/tags/simplejson-2.1.0/d...</a><p><a href="http://codespeak.net/lxml/lxmlhtml.html" rel="nofollow">http://codespeak.net/lxml/lxmlhtml.html</a>
<a href="http://codespeak.net/lxml/elementsoup.html" rel="nofollow">http://codespeak.net/lxml/elementsoup.html</a>
<a href="http://codespeak.net/lxml/html5parser.html" rel="nofollow">http://codespeak.net/lxml/html5parser.html</a><p><a href="https://launchpad.net/oursql" rel="nofollow">https://launchpad.net/oursql</a><p>EDIT: checked the date, the article is from 2008. I guess that explains things. :)
I wonder how large the proportion of social scientists is who actually know that Python is a programming language.
(I am a social scientist myself).<p>In my own experience, it is already pretty hard to teach them R.
For agent-based models -- what the author seems to enjoy -- also see <a href="http://cs.gmu.edu/~eclab/projects/mason/" rel="nofollow">http://cs.gmu.edu/~eclab/projects/mason/</a>
I would like to recommend my own project, GarlicSim: <a href="http://garlicsim.org" rel="nofollow">http://garlicsim.org</a><p>It is a Pythonic framework for simulations. It's still in alpha, and I'd be happy to help people start using it in research.
Glad this list still resonates with researchers.<p>I have added an addendum to the bottom of the list promoting the Enthought distribution (<a href="http://www.enthought.com/" rel="nofollow">http://www.enthought.com/</a>) of Python, which includes most (maybe all) of these packages in a single distribution.
This seems to be a good list for anyone using Python in their research, not just social scientists.<p>I used numpy for all my numerical analysis homework. The professor always lauded me for having the most concise code in the class.