Originally it was a Plone site but after a year we realized it was too hard to meet the deadlines of NASA VIPs. So we moved it to Django. Also, myself and the other developers have NEVER been happy with the flash on the front page. :P
The scariest bit of code was probably the scraper + html5lib normalizer I used to migrate something like 15 years of legacy static HTML content from two separate into the CMS using a large pile of selectors and cleanup heuristics.
http://chris.improbable.org/2009/10/16/deploying-django-site... describes how the site was packaged using RPM so we could fit within the more traditional IT model used by the production sysadmins.
One open-source tool which came out of this was https://github.com/shentonfreude/webcompare which we used to bulk-compare the old and new sites to confirm that all of our legacy URLs either redirected or were intentionally 404/410ed.
I have learned from some of your posts about developing with Django, so I took a second look at the site after realizing you had helped build it.