Multiple tags radio

Finally, last.fm allows me to listen to a radio station consisting of multiple tags (though there’s a maximum of three for some reason). Now I can listen to “Swedish indie rock with female vocalists” radio.

Singing ladies who are beautiful

To expand upon the posts here, here, and here, I’ve discovered a new band. I just discovered The Dø today and proceeded to buy a ticket for their concert at the Roundhouse in Camden next week. Thank you, last.fm, for recommending events to me. I love you.

Anyway, if you don’t think “On My Shoulders” and Olivia Merilahti are hauntingly beautiful, you are not human:

“At Last” is in a major key, which is more of my usual type of song, and the video is weird and cool and has oranges:

Be my friend on last.fm!

Felix Mendelssohn on tour

Sometimes I get confused when last.fm tells me the Felix Mendelssohn is “on tour”.

Tags are good, folders are bad

Organization by means of folders is bad for most things because it requires me to determine a hierarchy of categories, to determine the relationship among objects. Also, a traditional hierarchy only allows for an object to exist in only one place, i.e. with only one path which defines it.

Organization by means of tags is good because the organization is an emergent property, an epiphenomenon of the system based purely on semantic relationships, the “semantic network” of the system. Also, tags allow objects to have multiple “connotations”, if you will.

For example, last.fm successfully uses tags (well, I guess it’s the users of last.fm who assign tags) to help determine the relationship among artists, songs, albums, etc. In this way, The Cardigans can be tagged with “pop”, “rock”, “female vocalists”, “swedish”, “alternative”, etc. When tags are proposed by a sufficiently large (and sufficiently diverse) population, an artist’s place in the world can be more fully understood. To categorize The Cardigans under “pop/rock”, like CD stores do, is not only inaccurate but also insufficient. Once there is a great enough volume of tags, i.e. meta-information (information about artists, songs, etc.), the weighted semantic network of the system shows how closely related the tags we have chosen are.

Semantic networks are beautiful.