MyTV Genie
I use Windows Media Center as my PVR, and for the most part, like it a lot. It’s a huge step up from the PVR capability built into the cable box that Rogers supplies, although it can’t do HD.
Something simple like “record every new episode of a show” is not possible with the Scientific Atlanta PVR that Rogers supplies. It will either record every episode (which, with time-shifting channels, means you get way more than you wanted), or every episode in a particular time slot (which means you miss it if it moves).
TiVO does all this too, but requires a subscription for the data. I wouldn’t mind paying it if I could get an HD TiVO, but I don’t really want to pay a subscription for something I can get from Windows Media Center for free.
One TiVO feature I don’t get with MCE is predictive recording. This is the ability that the TiVO has to analyze what I’m watching, what other people are watching, and automatically record shows it thinks I might like.
With a PVR you spend less time flipping channels, so it’s harder to run across new shows you might like; predictive recording is a good way to find new stuff, assuming it works well. And it’s something that Windows Media Center doesn’t have built-in.
Enter MyTV Genie.
This is an MCE add-on that adds predictive recording to Media Center. I’ve been looking for this capability for a while, so I’m surprised I haven’t run across this before.
So far it’s only suggested a couple of shows to me, and they’re not ones I love, but I’m hoping that more users will equal better recommendations.
Our TV watching pattern must seem schizophrenic to a tool like this, since TV time here is divided into roughly three distinct categories: Toddler Time (Thomas the Tank Engine, Sesame Street), Wife Time (Dr. Phil, Mcleod’s Daughters), and Adult Time (Nanny 911, Flip This House, The Daily Show).
Either the prediction engine needs to find someone else that also watches TV in roughly this combination, or some Daily Show viewers are going to start getting Sesame Street recommended to them. It will be interesting to see what happens.
January 17th, 2008 at 12:58 pm
When determining what show to recommend, they could key off of a category index (another interesting key would be time spent watching a particular category). Just because you watch Friends, Seasame Street, and Flip this House, it doesn’t necessairily mean I’ll see Seasame Street in my recommended lineup (unless I watch Thomas the Train). If they didn’t try and match category with category, then it would be difficult to imagine anything other than a seemingly random assortment of recommendations.
Some really interesting results could show up, depending on how complex they make their viewing behavior model: user feedback, user/movie ratings, type of content being watched (action, adult, comedy, etc.), and even correlations between television stars and your most watched movies or shows. Imagine getting a recommendation for a movie or television show made ten years ago which features Bruce Campbell because I seem to be a big fan.