Once referred to as “tripletags” supposedly but now called Machine Tags, due to Flickr’s adoption of their use for geocoding, etc. From machinetags.org:

While still using the tag name, it’s slightly more controlled and has some interesting potential. While their main use seems to have been the geocoding your probably familiar with, others are beginning to expand it’s use to pull information from various API’s into their software or sites. From clagnut:

But what you’ll probably find more interesting is the discussion of book tagging:

Adactio has a nice follow-up article on the use of machine tags with flickr and blog posts, including the benefit of having custom namespaces:

There’s also a nice post over at geobloggers about some of the possibilities. And if your an RDF fan there’s the poor man’s RDF post you may find worth reading.

Those is libraries are probably familiar with namespaces and info uri’s. Maybe they’ll be submitted as a namespace on machinetags.org, who knows. Another similar use is that of “dctagging” by Pete Johnston (thanks to edsu for the link):

dctagged dc:creator=Berners-LeeTim

Presuming the API’s are decent, I can see quite a bit of potential for pulling together data from disparate sources for augmenting records and other data. A library’s flickr photos of an author’s visit can be easily shown on record pages in the catalog. A work with geo-location data could not only be mapped but also have photos and events from modern day. While there are other ways to accomplish the same thing, machine tags have the potential to make it somewhat easier and use a larger community. I think it would be interesting to see what data could be added to the view of WorldCat Identities if dublin core machine tags started being used. Photos, online texts, etc.

Just some thoughts, I haven’t gotten around to playing with them yet. And if you didn’t notice Google Maps now supports GeoRSS sources.