HereCast is "an open infrastructure for location-based services using Wifi". I.e. it allows users to determine their geographic location based on the wireless access points (AP) there computer can see. This is similar to what I proposed in my Linux Journal article on Geolocation.
However, this system does appear to require a user-built database of AP MAC addresses -> location. It is nice in that the system displays the canonical name for the location, rather than just a confusing latitude/longitude. However, such user-built systems require such a long-term investment of *potential users* before anyone can be an *actual user* they almost never work. The activation energy is too high to start the reaction.
Now, it people started embedding location information *inside* the AP, then the system could spread independent of a central server. It still requires activation energy, but does not require a single-point of information entry/retrieval like many of these other systems do.
If databases are still necessary, it would be better to develop a system where any number of "database servers" could gather information for a region (or just friends, a team, company, etc), and then publish this information via an RSS feed or some other mechanism that any user could then slowly gather and use as necessary.
"Oh, I'm heading to New York, let me grab a link to the NY-AP-LD (New York Access Point Location Database) and I can track my progress & look up useful info!