I honestly think this situation will improve since there is a lot of competition. It's not so much saying whose product is best, it's whose product is better suited. Competition will help deteriorate any of these problems.
netflixprize.com anyone? A good algorithm someone comes up with for this thing could probably be reapplied to the scenario you mention. The problem mentioned isn't exactly hard to master. However, the contexts in which such technologies would be used are very limited to a different user's tastes. You could make an algorithm that's perfect for yourself, but unless your program gathers extensive data from the user, it's just not going to be right for everyone. That limits the technology in the fact that it has to learn about the user rather than being ready to go.
The amount of memory it would take to analyze a huge amount of photos or music in complete accuracy seems impractical, so you would probably end up making mass generalizations for an efficient algorithm.
If you could analyze connections in the data, you could network songs or photos together. There's also the problem that a lot of music contains more than one mood, and some people might prefer their information sorted in different ways. One song or photo could be a part of several networks.
Then there is always the possibility of the program categorizing something in a way in which an individual person disagrees. The job of such technologies should be simply to assist the user, not to make things a rule. Meanwhile, it should respond to the users actions in itself. As well, there are many things that are dependent on context rather than solely recognition (ex. fair use vs. copyright infringement).
In these senses, I am personally unhappy with the fact that the GTK libraries will assume what is best for me as a user. I particularly have distaste for the GTK combobox in terms of visibility of information, which is more important to me than how far I have to move my mouse.
GNOME simply makes too many assumptions of its users and that's why I don't use it.
I praise KDE for being more configurable, but not for sensory overload on a new user. This can be improved, however. Meanwhile GNOME is handicapped in that it doesn't have a lot of functionality. You can't innovate on configuration usability in functionality that you don't have.
Slashdot Mirror
I honestly think this situation will improve since there is a lot of competition. It's not so much saying whose product is best, it's whose product is better suited. Competition will help deteriorate any of these problems.
netflixprize.com anyone? A good algorithm someone comes up with for this thing could probably be reapplied to the scenario you mention. The problem mentioned isn't exactly hard to master. However, the contexts in which such technologies would be used are very limited to a different user's tastes. You could make an algorithm that's perfect for yourself, but unless your program gathers extensive data from the user, it's just not going to be right for everyone. That limits the technology in the fact that it has to learn about the user rather than being ready to go. The amount of memory it would take to analyze a huge amount of photos or music in complete accuracy seems impractical, so you would probably end up making mass generalizations for an efficient algorithm. If you could analyze connections in the data, you could network songs or photos together. There's also the problem that a lot of music contains more than one mood, and some people might prefer their information sorted in different ways. One song or photo could be a part of several networks. Then there is always the possibility of the program categorizing something in a way in which an individual person disagrees. The job of such technologies should be simply to assist the user, not to make things a rule. Meanwhile, it should respond to the users actions in itself. As well, there are many things that are dependent on context rather than solely recognition (ex. fair use vs. copyright infringement). In these senses, I am personally unhappy with the fact that the GTK libraries will assume what is best for me as a user. I particularly have distaste for the GTK combobox in terms of visibility of information, which is more important to me than how far I have to move my mouse. GNOME simply makes too many assumptions of its users and that's why I don't use it. I praise KDE for being more configurable, but not for sensory overload on a new user. This can be improved, however. Meanwhile GNOME is handicapped in that it doesn't have a lot of functionality. You can't innovate on configuration usability in functionality that you don't have.