Scott VanDenPlas

I don’t know about everyone else, but I personally never use Wikipedia’s internally indexed search engine. Every time I have tried to search for something in the past, I rarely get meaningful results. For example, one major thing lacking is spelling correction and suggestion. If I want to look up the wikipedia article on Ghengis Khan and I misspell the query, external engines make the correction for me. Look at the results difference between Google and Wikipedia. Using Google, I find more accurate, useful, and contextual results than using Wikipedia’s internal engine. Wikipedia’s search is cumbersome, slow, and inaccurate.

I just finished watching the natural language search implemented by True Knowledge over at Tech Crunch. I think the results returned are brilliant in certain applications. For instance, if I want to read the a history of the civil war, that is not something easily summarized in a natural language response, so True Knowledge is not offering me anything more than a civilized way of querying their engine. However, when it comes to specific facts, both Wikipedia and True Knowledge have an incredible amount of overlap.

If I want to know how tall the Sears Tower is, I don’t _need_ to parse through an entire Wikipedia entry on the history and exact specifications of the building… I want to know that the Antenna/Spire is 527 m (1730 ft) and that the Roof is 442 m (1451 ft). I would bet that at least a fair amount of Wikipedia users are looking for simple facts, easily addressed with responses such as this.

Now if only I could have a reference search to go along with that or some sort of auto-verify with authoritative sources.

Here’s the True Knowledge video: