Semantic Mediawiki

A few of the talks I went to at wikimania were about semantic mediawiki, and wiktionaryZ. I am somewhat interested in wiktionary now, but I still have some reservations, I am however very interested in semantic mediawiki. This post is about how wikipedia can use it, a little later I will post another on how I think this makes mediawiki a lot more desirable from a small groups perspective also.

So, it’s just an extension to mediawiki, so installation is rather easy, although I’ll be able to tell how easy later or tomorrow probably… Basically it allows you to describe relationships between the current article and other articles, or just give the current article attributes.

For example, there may be an article on Canada, in the normal wikipedia you might say something like Canada is a country in [[North America]]., which would create a link to the article on “North America”. In semantic mediawiki you can instead say Canada is a country in [[located in::North America]]. This allows a relationship, which can then be queried, organized and exported via RDF (Resource Description Framework) (there are more powerful features but this is the basics). There are 2 huge benefits of this.

Right now on wikipedia there is fairly extensive replication of data, for example, the article on Baseball might list the major teams, but then there is also a list of teams, and an article on each team. If you change some information about a team on the team article it does not update anywhere else. Everything is just flat pages. Human editors have to go and make all the changes, which isn’t a huge problem, but it is a lot of work. Additionally, things like lists (for example) could be much more easily maintained if instead of a huge text block were actually query outputs.

The second huge benefit is external queries. Of course there is a huge amount of information in wikipedia, but right now the only way to get to it is to go read it. If you want to do anything complicated with the data you are pretty much out of luck. Think instead if you could query the data in a structured way. Maybe you want a list of all the countries in the world, their populations, and current leaders displayed on your geography site, with this you could. You could do mashups with the data, maybe see who is on the board of major companies, their relationships, and tie it in to google news and see news about them. There are obviously a lot more ideas, and since this is all GFDL people are free to create business models on any of them.

Anyway, it is all very exciting, more to come about what I personally want to do with this soon.