Wolfram recently launched their new service for information retrieval. The service is, as far I am concerned, not a search engine – but a fact engine. And a very clever fact engine.
I am deeply impressed by the service, probably due to my awe for anyone trying to overcome the barrier of artificial intelligence (or close to it) – I certainly strive at not embarking on AI:)
In short Wolfram|alpha lets you ask a question, or provide some keywords and in return you get a set of facts that (the engine believes) are relevant according to the input. This looks fairly similar to what Google provides; enter keywords and get relevant results. However, the huge difference is that Google only provides documents that it does not understand, and it provides results which is believed to have some relevance which they might or might not have. Wolfram on the other hand provides you either with facts (which the machine knows) or they provide you with nothing. So the precision of the result is bound to be better (i think). However they do not know everything, actually very little (in the large sense).
There is much headlines which essentially is of the kind; “Google competitor launched”, “Wolfram takes on Google” etc,etc.
Well, are they competitors? I think no. Wolfram|alpha is for me an advanced (dare to say intelligent?) dynamic encyclopedia. A place where you know more or less exactly what you want and you get an answer – or not. Google on the other hand not a place where you get answers, you might even not know what you want (!) – it provides a way to navigate an enormous information space – but does not give you any answers.
In my opinion the two services are orthogonal. Both are very solid information retrieval services, but also very different, which differentiate them and make place for both. But that’s my opinion. Are the two services (direct) competitors? Will both survive the harsh climate of the internet?
Geomatikkblog
I agree, the two engines are helpful in different situations where different approaches to knowledge matter.
I’d love to see some expansion in the crawling regime (if any?) and some serious contributions (http://www58.wolframalpha.com/participate/participate.html) to advance Wolfram. In its current form, I find it way too “arrogant”.
And by the way – http://en.akinator.com/ is a quite impressive little creature. What about a stepwise approach in search engines?
I agree that wolfram is quite ambitious and the front-end is quite arrogant at first try.
I too wonder why or how the crawling engine works, or the rational behind it. It seems that it actually crawls very little (or at a low level of intelligence). How hard can it be? ;)
Looking forward to expansion of the knowledge base, either by participatory means or by automatic crawling – one approach could be a combination i.e. semi-automatic knowledge construction. A very interesting field:)
Additionally they should make the system even more transparent, so for instance one could understand how the system works (high level), for instance to enable web developers to adapt their site appropriately and “ping” wolfram to crawl it. Would potentially boost semantic web a bit.