Paul Ramsey, at the time Refractions' president, was striving to finalize first release of GEOS, which would have made possible for PostGIS to perform all the operations in the OpenGIS Simple Features for SQL Specification. Contract's main focus was indeed reaching that target as soon as possible.
As a core developer I've been feeding
PostGIS during its 0.8 and 0.9 series
I became maintainer during the 1.0 transition (2005) and retained the role till January 2007 (1.2.1 version).
What happened while I was away (hacking Gnash full time during 2007-2008) can be summarized as follows:
By July 2009 release 1.4 hit the streets. This was much more exciting than previous release: prepared geometries, cascaded union, JSON output, new algorithms, improved testing framework, improved documentation, ... (see release notes). You could tell by 1.4 that PostGIS (it's community) was fully healthy again, probably due to improved (more open) governance model.
I got back to PostGIS hacking in January 2009 (with the WKTRaster adventure), during this "revitalization" process. A perfect time :)
This is an excellent example of how good of an investment
free software is: got away for 2 years and all the work I did
wasn't just still there, but people actually improved
it and built on and around it.
Also, the market knew me and most important the work I did, which isn't easy when all you do goes into proprietary software (who made Oracle Spatial ? and ArcGIS ? can you evaluate their work ?).