My last project at WikiVote was to create a knowledgebase for the main Russian news agency, ITAR-TASS. In the next posts I’m going to describe the project and analyse its results. This post is an introduction to the task we had.
For me it all started when I received a message from TASS telling how awesome the Semantic Web and semantic wikis are, and how much TASS needs these technologies. The head of computing in their monitor department have read my articles in habrahabr and got very excited about the dynamic way of creating the content and data semantic wikis provide.
I wondered, why does the news agency may need a wiki? How would they use it? Some time later I realized that the answer to this question is pretty straightforward: they will use it as the homemade Wikipedia, that is: they will store facts about people, events, places etc.
You may ask, “why not use the actual Wikipedia for that”? Well, they can’t do that because of the following reasons:
- Inaccuracy. The main news agency that represent Russia can’t rely on a crowdsourced encyclopedia. They’re afraid that the facts in Wikipedia will be inaccurate.
- Position. Of course both wikipedia and the news agencies try to remain as neutral as possible. Sometimes however, it’s hard to describe the events and things in a neutral way, so TASS can’t rely on a crowdsourced position of Wikipedia
- Proper accents. For example how are we going to describe Arnold Schwarzenegger – as a politics, bodybuilder or actor?
Besides I did my best to explain that non-semantic wikis sucks when the amount of facts is big enough and there is not a lot of editors working on the content.
So, the project began. Our system ment to replace this knowledgebase:
Yeah, there is a whole floor in TASS full of these card indexes. The smell of the paper and dust returned be to my childhood when I’ve spent days in the public library where my mother worked. The API of the knowledge base is the telephone and the old lady who know how the indexes are organized.