The last PI website rebuild was back in 2004 when we devised what felt like quite an innovative way of cross referencing content. But as is the way with our line of work, what seemed cool and groovy in 2004 had become pretty clunky and out-moded by 2010 (or probably even sooner for web fashionistas). And so the challenge 6 years on, was to refit the ship without losing any of its precious cargo and at the same time get that cargo out of the hold so that people could find it, read it and start taking action on it. And to do all that on very limited resources - of course. We used Drupal for the content managment but brought in some extra heavy lifting equipment in the form of an Apache Solr search for the content finding side of things. Solr is an incredibly fast and powerful search with its own database that is particularly well suited to handling queries on big archives like PI's.
We managed to keep costs down by using one of Drupal's ready-made page designs. And the over-busy twitterers at PI are managing to keep their web workload down with a nifty tool we built in that feeds their tweets directly into the system.
With the data migration successfully completed the new ship is now on her way again - fighting to secure our rights to a future free from government surveillance and interference .