Discover Music through The Music Genome Project by Pandora
There are two analogies in this post! What great value.
Together we set out to capture the essence of music at the most fundamental level. We ended up assembling literally hundreds of musical attributes or "genes" into a very large Music Genome. Taken together these genes capture the unique and magical musical identity of a song - everything from melody, harmony and rhythm, to instrumentation, orchestration, arrangement, lyrics, and of course the rich world of singing and vocal harmony. It's not about what a band looks like, or what genre they supposedly belong to, or about who buys their records - it's about what each individual song sounds like.
An interesting metaphor applying some popular concepts of the genome to music. There is nothing particularly surprising or unusual about this but I find it an interesting example of a metaphor that illustrates some of the folk theories present in both the sources and target domains.
Hear more about this on the Inside the Net 6 where The founder makes an interesting comparison to Amazon or iTunes suggestion services which are not based on expert classification but rather the collective wisdom and popularity of users. It would also be interesting to compare this very labor intensive human classification project with the latest service in Winamp that offers suggestions based on the actual sound. This, if of course, reminiscent of some of the debates within linguistics, for instance, regarding corpus tagging where the question whether to use a stochastic or rule-based parser was ultimately resolved eclectically (with a slight edge to the stochastic side which can simulate bootstrapping). The same goes for speech recognition, machine translation, etc.
Technology note: But metaphors and bootstrapping aside, the streaming radio on Pandora.com using the Genome project is great value (both the free and commercial portions) despite its decision to use a flash based interface which while fast and reasonably convenient is not very intuitive and breaks lots of open standards. Hopefully, Ajax is just around the corner (but that may be just wishful thinking).
Other services such as Accuradio or Music Match are decent alternatives but so far, Pandora seems to offer a bit more. What it's missing (partly due to the interface) are links to more information about artists (Music Match's stron point) and more convenient links to Amazon (what I like on Accuradio).
Life note: The main stated advantage of Pandora is its supposed ability to open access to new music based on real preferences rather than the commercial wishes of music companies. In a way, the machine is replacing the live DJ and being more successful because it is unencumbered by compromises. The ability to share one's own stations probably contributes further value. However, what is missing is the element of conversation with somebody actively expressing themselves through their conscious choice. (I'm not yet sure whether this is good or bad. Leaning toward a neutral trade off). The (possibly) interesting part is the fact, that due to automated playlisting in most commercial radio stations, a machine can now offer more of a human touch than humans. But the Turing-test-type question is whether it is more important to us to have the feeling of human contact (ie. a live presenter linking the songs, even if they, unbeknownst to us, were selected by a computer based on commercial considerations) rather than creative flexibility and openness eventhough they are offered to us ostensibly by a computer which, however, is basing its decisions based on artistic judgements of real humans.