Semantic MediaWiki: The code is out there

Finally! 500 nice lines of code, including the AJAX-powered search, and that's it, version 0.1 of the SeMediaWiki project! Go to Sourceforge and grab the source! Test it! Tell us about the bugs you found, and start developing your own ideas. Create your own Semantic Wiki right now, today.

Well, yes, sure, there is a hell of a lot left to do. Like a proper triplestore connecting to the Wiki. Or a RDF-serialization. But hey, there's something you can play with.

Semantic MediaWiki Demo

Yeah! Doccheck's Klaus Lassleben is implementing the Semantic MediaWiki, and there's a version of it running for quite some time already, but some bugs had to be killed. Now, go and take a look! It's great.

Docchecks Semantic Wiki Demo

And the coolest thing is the search. Just start typing the relation, and it gives you an autoexpansion, just like Google Suggest does (well, a tiny bit better :) Sure, the autoexpansion is no scientific breakthrough, but it's a pretty darn cool feature.

The SourceForge project Semediawiki is already up and running, and I sure hope that Mr Lassleben will commit the code any day soon!

Even better, Sudarshan has already started implementing extensions to it - without having the code! That's some dedication. His demo is running here, and shows how the typed links may be hidden from the source text of the wiki, for those users who don't like it. Great.

Now, go and check the demo!

New people at Yahoo and Google

Vint Cerf starts working at Google, Dave Becket moves to Yahoo. Both like the Semantic Web (Vint said so in a German interview with c't, and I probably don't have to remind you about Daves accomplishments).

I'm sure, Yahoo got Dave because of his knowledge about the Semantic Web. And I wonder if Google got Vint out of the same reason? Somehow, I doubt it.

Another Semantic MediaWiki

I stumbled about another Semantic MediaWiki, an implementation created by Hideaki Takeda and Muljadi Hendry of the Japanese National Institute of Informatics in Tokyo. Their implementation looks very neat, although it is quite different in a few basic things (that we consider crucial in order to work), take a look at their full paper (it's in their wiki - oh, and it's in Japanese).

The basic difference between their approach and the one we suggest is that they add metadata management abilities to MediaWiki - which is cool. But they don't seem to aim at a full integration into the Wikipedia, i.e. embedding the metadata into the article text instead of appending it at some place. Actually, if we had software that is able to process natural languages, we wouldn't need our approach, but their would still be useful.

Nevertheless, they have a big huge advantage: a running system. Go there, take a look, it's cool! Actually, we have a system online too, but won't yet disclose the link due to a bug that's a kind of showstopper. But expect it to be online next week - including the source and all! It will be just a first version, but I sure hope to gather the people who want to work on it around the code.

Commited to the Big S

Not everyone likes our proposal for the Semantic Wikipedia. That's not a big surprise really. Boris Mann was talking about the advantages of tagging, and some ideas like blessed tags, that sounded very nice, when Jay Fienberg pointed him to the Semantic MediaWiki proposal. Boris answers: "I notice with a shudder however, that the Mediawiki stuff uses a large "S" Semantic, and includes RDF. I admit it, I'm afraid of RDF."

Yes, we do. And we're proud of it. Actually, it's the base for the better half of the possible applications we describe. Jay has some nice answers to it: "I think the MediaWiki folks are just recognizing the connection between their "tags" and the big "S" Semantic Web [you bet!, denny]. There are taxonomies and ontologies behind the popular tagging apps too--folks behind them just aren't recognizing / publicizing this (for a number of reasons, including that tags are often part of a practical application without big "S" Semantic Web goals). [...] I'm not a super huge fan of RDF myself, but I think it's useful to not be afraid of it, because some interesting things will come out of it at some point."

Our idea was to allow the user to use Semantic Web technologies even without really understanding them. No one needs to understand RDF fully, or OWL, to be able to use it. Sure, if she does, well, it surely will help her. Any by the way, RDF really is not complicated at all, it just has a syntax that sucks. So what?

Maybe it's a crude joke of history to start the Semantic Web with syntactic problems...

By the way, does anyone have a spare invitation to GMail for me? I'd really like to check out their service. Thanks, Peter, that was fast.

Semantic Wikipedia

Marrying Wikipedia and the Semantic Web in Six Easy Steps - that was the title of the WikiMania presentation we gave about a month ago. On the Meta-Wikipedia we - especially Markus Krötsch - were quite active on the Semantic MediaWiki project, changing and expanding our plans. DocCheck is working right now on a basic implementation of the ideas - they have lots of Wiki-Experience already, with Flexicon, a MediaWiki-based medical lexicon. We surely hope the prototype will be up and running soon!

Wow, the project seems perceived pretty well.
Tim Finin, Professor in Maryland: "I think this is an exciting project with a lot of potential. Wikipedia, for example, is marvelously successful and has made us all smarter. I’d like my software agents to have a Wikipedia of their own, one they can use to get the knowledge they need and to which they can (eventually) contribute." - Wikipedia meets the Semantic Web, Ebiquity blog at UMBC

Mike Linksvayer, CTO of Creative Commons: "The Semantic MediaWiki proposal looks really promising. Anyone who knows how to edit Wikipedia articles should find the syntax simple and usable. All that fantastic data, unlocked. (I’ve been meaning to write on post on why explicit metadata is democratic.) Wikipedia database dump downloads will skyrocket." - Annotating Wikipedia, Mike Linksvayers Blog

Danny Ayers, one of the developers of Atom and Author of Atom and RSS Programming: "The plan looks very well thought out and quite a pile of related information has been gathered. I expect most folks that have looked at doing an RDF-backed Wiki would come to the same conclusion I did (cf. stiki) - it’s easy to do, but difficult to do well. But this effort looks like it should be the one." - Wikipedia Bits, Danny Ayers, Raw Blog

Lambert Heller of the University of Münster wrote a German blog entry on the netbib weblog, predicting world domination. Rita Nieland has a Dutch entry on her blog, calling us heroes - if we succeed. And on Blog posible Alejandro Gonzalo Bravo García has written a great Spanish entry, saying it all: the web is moving, and at great speed!

So, the idea seems catching like a cold in rainy weather, we really hope the implementation will soon be there. If you're interested in contributing - either ideas or implementation - join our effort! Write us!