Collaborative filters & playlists

Been compiling playlists using iTunes and thinking what this aggregation means. Here are some ideas and questions.

Do my music preferences reflect memories more than interests?
If others like my song choices does this mean we share other common interests (aside from the music) or do we share a common time-frame? - we were listening to the radio during a comparable formative period (12-22 years old?). Will my music buddies share my interest in KM or my curiosity for Internet collaboration? - uncertain.

Does a shared music taste provide a sufficient introduction and 'hook' to lead to deeper sharing?
Often wonder if on-line affinity depends on a broad interest profile match or if strongly shared preferences for just books, only music tastes or photo images is able to jump start future collaboration, exploration and learning?

What is the exact role of a common goal vs. a shared genre taste, contributing to on-line work comfort?
On-line compatibility seems to be a strange mix of similar purpose, affinity for written or audio clues and language / tone attractions, while trust, social capital and reciprocity tend to muddy the waters here.

Interesting to see tools for playlists starting to emerge, compare their qualities and features. David Jennings has a useful table.

Playlist comparison matrix

Feedback, comments, ratings, profiles enhance the community that is developing around swapping playlists. Links to blogs and music portals (iTunes, Odeo), download, uploads, exports, sequencing, ripping and mixing seem to be common functions. A growing need is for portability across spaces, players and PCs.

I'm still uncertain what a playlist really represents, how much it tells us about the owner, the exact role a playlist has in defining other qualities and the position a playlist will come to play in personal digital portfolios.

Any ideas to share?

Comments

[Rup3rt]

Recommendation versus AI. Collaboration versus Intuition.
Seen Pandora yet?

http://dekstop.de/weblog/2005/08/a_first_look_at_pandora/

[David Locke]

Hook to learning depends on whether you are trying to learn to dance, sing, or play some musical instrument. You could use a playlist to order the content you are delivering.

If I understood this correctly, podcasting isn't just about music, so playlists don't have to be exclusively about music either.

[Bryan Alexander]

Fine post, Denham.

Memories: add to that the way playlists reflect the experience of life while using iTunes, a time period which extends, and therefore becomes another layer of memory. "Oh yeah, that band was *so* 2004."

Hook for learning: I don't think so, not without other connections, better tools. Seb Paquet has been looking into Webjay.

[David Locke]

Stewart Brand tells us in his "How Buildings Learn" that remodeling is an act of embedding our implicit fitness into the physical artifact of the building. We can't know what it really means to put a hole in the wall, but we know that we are improving the fitness of the building towards ourselves.

Radio stations and other media businesses us content to draw audience and audience to define the station. Channel sellers and retailers do the same thing. Inventory is them, inventory is us.

Artifacts accumulate a community and a culture depending on the depth of individual subscription to the underlying culture.

Interest in hobbies typically last two years, but we know people who have spent a lifetime in a hobbie. The reality is that there is a statistical distribution describing a hobbie or a culture. They tend to be plossion distributions. Plossion games are descritive of culture as is the Long tail.

In Argentine tango communities a preference gets built up between those that can dance to Piazolla and those that can't. Those that dance to neuve, and those that dance to traditional. These are paradigms. They are not exactly mutually exclusive as some claim that cultures are. But, paradigms do represent the contours of a cultural typology.

Country music is likewise: tradition, outlaw, new, rap-infused. I'm moving away from it these days. I'd rather listen to 60's - 70's oldies right now. These are generational, while the country music or hobbies might last for periods shorter or longer than the generational eight years.

Retailiers can use data mining against your purchase histories and tell you things, but they don't focus on individuals. Personalization software trys. But, how often is that stuff correct. What I bought in the past is only a partial indicator of who I was when I bought it. I may not be an indicator of who I am today.

Your playlist will tell on you. What it can tell is up to the analyst. It is implicit. One day it might be explicit.

It's like a conceptual model as well. How does your conceptual model for a given domain differ from that of other people's. The diffrence will look like a Plossion game. They are asymetrical. A Nash equalibrium won't hold.

I read somewhere that cultures are exclusive. I don't think so. We have so many attributes to associate around. There are so many artifacts to associate around. We manage so many things at one time, but we also multitask around risk and subscribed identities. So our playlist will reflect the subscribed identities at the time something was played. It will represent our opportunities for cultural migration, and out cultural commitments in terms of am I just a know what, a know who, or a know how--am I an expert--would I spend tons of time trying to find the sheet music for "Teach me to Dance."

The same thing is true of definitions like the definition of KM. We subscribe to a definition. We commit to a definition. So our definition says volumes about us like our playlists.