My music communities

Among my research interests are tool systems. I want to know how (communities of) users combine different tools to support complex workflows, often leading to usage patterns very different from those originally intended by the tool designers. Earlier, I gave an example of how we used a combination of tools to help us in the difficult collaborative process of writing a call for papers.

A very different example is how I am using three different tools to build a collection of my favourite songs on the Web. Already for quite a while I have been a subscriber to Last.Fm, the “social music revolution” site. This site allows you to enter an artist, tag or username. A music stream is then dynamically generated, often starting with a song from, for example, the artist you entered, to be followed by many other songs from related artists.

Last.Fm is a great way to learn about new music, so it builds my list of interesting artists. It does not help you to retrieve that one song you are craving to hear right at this moment, however. Other tools are better suited for this, one of them being Songza. This tool allows you, Google style, to just enter a string and then gives you all the versions of songs of which the artist name or song title match the string (It is also great for parties: last Saturday, I had some friends over to play a board game. We had my mini laptop next to us on the table, and in between making moves on the board, we took turns selecting our favourite songs, together acting as a kind of “distributed DJ”)!

One nice feature of Songza is that you can save a playlist (it would be much better to be able to save multiple playlists, but as the tool is still being developed, this may soon be realized). Unfortunately, I am one of those sorry souls who have a notoriously bad memory for song titles, so how am I to fill that playlist with the songs that have touched my heart?

Here, another tool has come to the rescue. Actually, it is a regular website, but I informally define any piece of functionality that serves an individual or community purpose as a tool (see also the article I wrote with Mark Aakhus: Argumentation Support - From Technologies to Tools). The website is the Dutch National Top 2000 site. It contains the top 2000 songs as voted by listeners from all over the country, and is broadcast at the end of every year. My recipe was to browse the songs, select the ones that triggered a remnant of a musical memory, then looked up the title on Songza, and added it to my playlist there.  The result, an (emerging) playlist of my golden oldie favourites.

Interesting from a community perspective is that, even though my playlist is about as personal as it gets (I don’t blame you if you never want to talk to me anymore having seen some of my favourites :-)), it could only have been generated by the efforts and intersection of three huge online music communities. They not only supply the content (Last.Fm and Songza), but also the relevance measures for that content through community members discussing and voting for their favourites.

It is fascinating to ponder how all these different individual and community levels paradoxically mix and mingle to provide me with my ultimate personal listening experience. If this isn't a telling example of how every person is shaped by his community (or rather, communities)...

Comments

[CLK]

Interesting to see how the boundaries between the concepts of tool and community are not that clear in the online world. Members of a community use a tool to interact, carry out a process. Tools posses certain functionality which allow users to connect between them and perhaps form communities.

I am thinking, as a developer how would I know when a certain tool or its functionality will encourage the creation of a community?

[Aldo de Moor]

Good question. Very hard to know from the technical properties of the tool themselves. Many tools play very different roles in communities than the ones intended by their designers. Tools are often also used in concert, together providing a subtle mix of community functionality. The best way to find out are empirical studies of tools-in-use and then try to generalize from there.