Do you intend to analyze the users taste/preferences?
This is just a suggestion and in all probability you guys must have thought about this. So I'm sorry if I am wasting your time.
Why don't you make a desktop application like last.fm that analyzes the music lying on a user's hard-drive using local resources and then uploads the info to you? I think that it may be possible for you to take a speech recognition API and modify it to recognize patterns in songs instead. Something like Pandora's music genome but automated, and then as User data builds up you can use it to predict what the user would like on your end.
Thus, this way you save server load by using the user's PC for some of that heavy lifting (i.e. analyzing their library) and analyzing the music on your server using your scheme would be a one off effort. You could make something like custom info files denoting details about a song that can be picked up when you run the app. It would be so cool to have a service like this.
P.S. - Thanks to you guys I am now in love with Yael Naim.
If you do that, make it optional. I kind of like just 'Nexting' through a random list, since I feel like it is exposing me to music and genre's it wouldn't had it known my current playlists.
Also, a search feature for me to find Yael Naim on mixest would be nice too :).
Otherwise, so far so good. Like the simplicity. Question: Is the only thing that 'More Obscure' does is to suppress songs you've heard before or don't want to hear again? The words imply a little more.
Only complaint is the registration password field needs to be obscured, instead of plaintext. Always disconcerting to type my password and actually see it, even in the privacy of my home PC.
Ah, so it sounds like 'More Obscure' means 'Make this artist more obscure in the playlist'? I interpreted it as 'drill down into even more obscure, less known, less listened-to artists'.
This is just a suggestion and in all probability you guys must have thought about this. So I'm sorry if I am wasting your time.
Why don't you make a desktop application like last.fm that analyzes the music lying on a user's hard-drive using local resources and then uploads the info to you? I think that it may be possible for you to take a speech recognition API and modify it to recognize patterns in songs instead. Something like Pandora's music genome but automated, and then as User data builds up you can use it to predict what the user would like on your end.
Thus, this way you save server load by using the user's PC for some of that heavy lifting (i.e. analyzing their library) and analyzing the music on your server using your scheme would be a one off effort. You could make something like custom info files denoting details about a song that can be picked up when you run the app. It would be so cool to have a service like this.
P.S. - Thanks to you guys I am now in love with Yael Naim.