Adam Krebs
09-14-2008, 06:00 AM
<div class='os_post_top_link'><a href='http://blog.wired.com/music/2008/09/zune-recommenda.html' target='_blank'>http://blog.wired.com/music/2008/09...recommenda.html</a><br /><br /></div><p>"<em>Microsoft showed us a sneak preview of the Zune 3.0 software it plans to release on Sept. 16 with the latest generation of Zune devices, and what we saw made iTunes' simple <a href="http://blog.wired.com/music/2008/09/steve-jobs-anno.html">Genius feature</a> look like a blast from digital music's past. While iTunes serves up a text list of recommended songs within your library and from the iTunes store, adding to the more basic recommendations its MiniStore feature used to make, Zune reinvented the recommendation concept by collapsing artists, albums and fans into the same recommendation engine, more accurately mirroring the way people think about music. The new feature, called MixView (pictured [below]), displays a single album, artist or user in the center of the screen and surrounds it with related items in a graphical format. ... You can start on an artist and instantly discover which bands influenced that artist and vice versa, by mousing over those surrounding elements in MixView. Double-clicking through to any song plays a 30-second sample, offers a chance to buy the track or, if you're a Zune Pass subscriber, plays the track in its entirety.</em>"</p><p><img src="http://images.thoughtsmedia.com/resizer/thumbs/size/600/zt/auto/1221368872.usr495.jpg" /></p><p>I feel that both approaches have their advantages. MixView, since it correlates the music <em>you actually play</em> with other peoples' recent listens, will provide a more current selection of music, showing what's hot <span style="text-decoration: underline;">now</span>. Also, it's less likely to correlate a song you've never played--for example, I have a large number of songs in my library I've never heard, most of which are utter crap and are therefore songs I don't want to hear more of. With Genius, your entire library is scanned and sent to Apple, matching the nodes against other people with similar <em>libraries</em> to you. This approach is better for songs you haven't heard in a while or downloaded in batch. My roommate, for example, que'd up "This Town" by O.A.R. and Genius returned "Camilo" by State Radio, a band who opened for Dave Matthews Band this summer (Dave Matthews Band and O.A.R. are very similar stylistically).</p><p>The problem with the Genius approach is that it can produce some odd results--even during the unveiling keynote on Thursday, when asked to show matching artists to Green Day, Genius <a href="http://twitter.com/gadgetgy/statuses/915346830" target="_blank">returned</a> Jimi Hendrix and John Mayer. We'll have to wait until the Zune 3.0 software is released to make a valid comparison, but I can only hope the algorithms evolve as both technologies mature.</p>