I find human opinion fascinating, and look forward to the glorious future when we can do better than up/down arrows.
Looking at a site like Rotten Tomatoes, where the metric is "professional reviewer thought thing was at least decent" and seeing that there is surprisingly little agreement even on such a simple metric (even a terrible movie, like Season Of The Witch, has a 10 so 10% of the reviews like it; even a great movie, like wall-e, only has a 94 - so somebody didn't like it - and most movies are in between with lots of critics liking and lots of critics disliking the exact same thing. Yet we still try to arrive at a truth by considering all their opinions equally as a group. This is obviously doomed to failure.)
It's even more clear when you think about something more abstract, like a favorite color. My favorite color is purple. If yours is yellow, does that mean I'm wrong? Clearly there isn't an obvious objective "This Is Good" when it comes to taste or opinions. Why then is up/down arrows the current state of the art?
I know Netflix has managed to do more by looking for similar reviewers and weighting their opinions, and Pandora does audio classification, but I'm not aware of this concept catching on well in other areas, and I particularly don't know of it being applied to multiple things at once - if I want a recommendation about a new movie that isn't on Netflix yet, or if I'm tired of the repetitiveness of Pandora's stations, their clever algorithms don't help me a bit beyond their limited walls. I don't know of a site like reddit or hacker news or videosift trying to capture more than 'good/bad' for evaluating links on the web. I don't know of a site that is really trying to help me pair my taste with other reviewers, so that I can get personally curated content and we can end up with a world that recognizes some people love action movies and others love romantic comedies and recommending a movie for me to see should take that into account. How wonderful will it be when there is finally a site that can suggest I should read these three articles because -I- will find them good, maybe I should be sure to catch Pacific Rim, and a new band I might like has a video out; while giving completely different recommendations to my sister, without trying to limit to an objective non-relative definition of "good", because we're talking about taste and opinions aren't wrong.
When one of you smart folks makes that site, please let me know.
There are some things that are so oddly polarizing that it becomes very difficult to guess whether someone will like it or hate it, no matter how much you know about that person. And even the people who are asked why they love/hate it tend to have a hard time explaining it.
I too look forward to this site. The problem with Smart Recommendations that Netflix and Pandora make is they are still subject to highly subjective tastes. I could have watched 6 documentaries on Wine, rated them all 5 stars and I still might not like the title recommended by Netflix, even if it is a documentary about wine and has an average 4.9 from other Netflix users. There are simply too many variables and most people (myself included) are fickle. If I'm in a certain I could like a movie I would otherwise dislike.
That would be hunch.com. It was bought by eBay a while ago, not really sure what's been going on with it. The concept is great, but in practice, your mileage may vary. There are a few other similar sites too... but none of them have really taken off in a big way.
unfortunately the fear of a company knowing too much in order to "sell" you, I.e. the privacy debate that is so rightly ongoing would restrict or otherwise cloud ambitions in this arena I fear.
Looking at a site like Rotten Tomatoes, where the metric is "professional reviewer thought thing was at least decent" and seeing that there is surprisingly little agreement even on such a simple metric (even a terrible movie, like Season Of The Witch, has a 10 so 10% of the reviews like it; even a great movie, like wall-e, only has a 94 - so somebody didn't like it - and most movies are in between with lots of critics liking and lots of critics disliking the exact same thing. Yet we still try to arrive at a truth by considering all their opinions equally as a group. This is obviously doomed to failure.)
It's even more clear when you think about something more abstract, like a favorite color. My favorite color is purple. If yours is yellow, does that mean I'm wrong? Clearly there isn't an obvious objective "This Is Good" when it comes to taste or opinions. Why then is up/down arrows the current state of the art?
I know Netflix has managed to do more by looking for similar reviewers and weighting their opinions, and Pandora does audio classification, but I'm not aware of this concept catching on well in other areas, and I particularly don't know of it being applied to multiple things at once - if I want a recommendation about a new movie that isn't on Netflix yet, or if I'm tired of the repetitiveness of Pandora's stations, their clever algorithms don't help me a bit beyond their limited walls. I don't know of a site like reddit or hacker news or videosift trying to capture more than 'good/bad' for evaluating links on the web. I don't know of a site that is really trying to help me pair my taste with other reviewers, so that I can get personally curated content and we can end up with a world that recognizes some people love action movies and others love romantic comedies and recommending a movie for me to see should take that into account. How wonderful will it be when there is finally a site that can suggest I should read these three articles because -I- will find them good, maybe I should be sure to catch Pacific Rim, and a new band I might like has a video out; while giving completely different recommendations to my sister, without trying to limit to an objective non-relative definition of "good", because we're talking about taste and opinions aren't wrong.
When one of you smart folks makes that site, please let me know.