Tuesday, October 16, 2007

Tufts 10K


Here I am with my running gal pals at the start of the Tufts 10K. We took the train to Boston, walked from Back Bay Station to the Boston Common, picked up our registration packets, and checked our bags. My fam fam caught up with us and got a picture of us before the race. Notice the dry hair.



Over 5 miles later, you can see that things have changed. I now look like a drowned rat because as soon as we hit the starting line, it started to drizzle. Moments after we crossed the finish line, the drizzle stopped.

My running buddy, K, and I ran the race together, from start to finish. She's the one who was with me when I collapsed back in May. Crossing the finish line together was an emotional moment. I felt very thankful to be able to participate in the race, to be able to run with my best time yet on that race, to share that moment with her, to have other women from Foxboro who joined us, and to have my friends and family there to cheer me on.

The Tufts 10K is a powerful race. You are surrounded by 7,000 other women, each of whom has her own story and her own reason for running. Because of the turns in the course, you have many opportunities to cheer on women who are both faster and slower than you. Families line up along Commonwealth Avenue, cheering their runners and holding up signs. While I won't deny the seriousness of what happened to me, I know that there were many women for whom this was a bigger step. It's both humbling and encouraging to think about all the stories that were out there that day.

Considering where I was in May, I am proud to say that I finished the race and shaved 8 minutes off my time compared to 2 years ago (results are here). I'll never be a fast runner, but I am certainly a thankful runner.

I hope I can do it again next year--maybe with a lot less drama!

Monday, October 1, 2007

Initial Thoughts about Musical Preference Systems

I’m a fan, and loosely a scholar of, Hawaiian music. So today I entered Keali’i Reichel, a well-known pop music artist in Hawai’i, into Pandora’s search engine. Pandora asked if I wanted something similar to “Keali’i Reichel (Holiday).” I said yes. It played a song from Hall and Oates “No Child Should Ever Cry on Christmas Day.” I gave it a thumbs down. The next selection was “Songs from Matengo Folktales” by Bowling Green State University Men’s Chorus. The next several selections were Christmas songs; apparently Pandora locked onto the “holiday” theme. Next I tried a searching for Israel Kamakawiwo’ole (a.k.a. IZ), arguably one of the best-known Hawaiian artists outside of the islands—it couldn’t find him.

Finally I tried searching on the term “Rainbow” hoping to find IZ’s rendition of “Somewhere Over the Rainbow/What a Wonderful World.” Eventually I saw something by Jake Shimabukuro, a Hawaiian artist. The helpful start message described the piece as featuring folk influences, mild rhythmic syncopations, acoustic guitar riffs, and “a good dose of guitar pickin’.” Not a bad description, I guess, for a ukulele piece. Perhaps the title of the piece, “When My Guitar Gently Weeps” fooled them.

Since I indicated that I liked this piece, Pandora used that as a model for choosing pieces for me. The next selection was “I Surrender All” from 50 Hymns on the Guitar. I guess it was a good dose of guitar pickin’. Subsequent selections included wonderful acoustic guitar pieces by such artists as Pat Kelley and Mark Holizinger, Don Ross, Raina Rose, and other guitartists with whom I was not familiar—but not a single piece of Hawaiian music other than the original Shimabukuro one.

I take this as evidence that programmatic musical preference determination must take into account more than statistical comparisons. I’m still thinking a lot about this, but my main area of concern is that the sense of community around a particular artist or musical style and the concept of identity through music play a large role in musical preference. As a further demonstration, consider a female listener who identifies herself as a lesbian. She likes the Indigo Girls. Based on musical characteristics (folk influences and a good dose of guitar pickin’), would Pandora select a Melissa Etheridge piece (basic rock song structures, extensive vamping, major key tonality, gravelly female voice)? Probably not initially, but to that listener the connection between these artists is real and important. There is considerable overlap in the communities that surround them since all of these women are openly lesbian musicians.

Elements such as community and identify are hard to quantify, but music preference systems will have to come to grips with the sociological and cultural aspects soon.