Dunno. Are tags that great as an organizational component in large installations? With hundred of thousand pages to maintain?
And that “fork discussions” thing sounds an awful lot like non-linear bbs’es (you know, where discussions are displayed indented).
In my opinion… yes… it’s the future, it’s the same concept as Nepomuk http://nepomuk.kde.org/node/1, it’s the same concept as Windows 7 special folders etc. It’s MUCH more flexible.
Well… not really… however that concept wasn’t so awful, just its appearance was awful.
This is just some brainstorming, though.
I’m really not convinced. Tags tend to add a lot of noise. They are great for small to midsize installations/projects. But I seriously doubt they’re a perfect solution.
Care to give a little more detail? Forking code is quite easy as it’s (most of the time) consisting of “materialized” thoughts. Forking ideas and discussions adds a huge layer of abstraction on that. How would that work?
And don’t let that sound too critical. By now I’ve just seen too much “usenet with a web 2.0 finish” rather than any real next “http with lipstick” solutions 
I disagree… a hierarchical organization is extremely limited… i.e. I have category->discussions, while using tags I can have the same discussion tagged with multiple “categories”, you can build a network between different items (discussions). Hierarchical systems are obsolete but they are still here because they’re what people is accustomed to. Microsoft understood this and they tried to develop the WinFS filesystem, the project was abandoned but they reused some ideas for things like Windows Vista and Windows 7 special folders.
Nepomuk is trying to bring something similar to KDE:[i]
[/i]As long as there have been computers there has been a very technical way of looking at information. Data is always organized in fixed hierarchies of folders or into tables. This is what slow computers back in the day could handle and which was always easy to program for.
This system worked quite well, likely because it was the only system users knew and it was somehow modelled on the way we organize papers in folders in our office cupboards. However, it is not at all how the human brain manages information.
The human mind depends a lot on context and relations. Nepomuk aims to make the computer work more like the human mind. (Of course we do not claim that Nepomuk is Brain 2.0, that would not be feasible today. But we can take inspiration from the way information is handled in the brain and use it to our advantage.)
In my opinion a perfect file-system would be a file-system without directories, where files are grouped through tags.
This concept also applies to a discussion system like a bulletin board.
You could fork a discussion so that you can talk about a sub-problem without hindering the main conversation, when the sub-problem has been worked out, you can propose to merge the discussion back to the main conversation… or you can develop that discussion into something which is totally different… just like code on GitHub, but you have messages instead of code fragments.
[b]
However[/b], as I’ve already said, this is only some brainstorming, I’m not going to build something like this in the short. These are just some ideas… the only pivotal concept is the tag-based system, I would never build another obsolete hierarchical system.