[wiki-research] Loaded questions about Semantic Wikis

Dirk Riehle dirk at riehle.org
Thu Jul 13 13:52:54 CEST 2006

>The hard part (and which, as CS points out seems to be mostly 
>ignored), is how to get people to write metadata that will actually 
>lend itself to automatic inference by a machine.

I agree---semantic wikis are missing a killer app that is convincing 
to those who are supposed to do the work. I'd love to see Wikipedia 
semantified, but who is going to do it and why would they? The 
benefits need to be tangible to those who are going to do the work, 
not some other party.

>How do you come up with a metadata scheme that is rich enough to be 
>able to describe all the important semantics of a wiki page? How do 
>you get people to agree on that when the industry can't even agree 
>on a standard for JavaScript?

That's what I don't worry about. For example, Wikipedia users are 
already doing a lot of the semantics stuff using categories. It is 
adhoc, chaotic, inefficient. But that's how things have worked before 
as well; it is an arduous process, but Wikipedia has proved that 
(some) order can emerge from chaos.


More information about the wiki-research mailing list