In the last blog I talked about Wolfram|Alpha. I just noticed that Google has something similar yet completely different in store: Google Squared.

So how is it similar? Both 'search engine' do not provide a list of results but try to provide a list of facts. But the way they produce these is very different and so are the results:

Google searches the internet (what else) and tries to extract semantic information from it. Resulting in sometimes vague, sometimes plain wrong results. Look for example at the color column in this square. According to it Ethanol is red. If you click on that attribute you can see the source and get an idea how google saw the words ethanol close to the word red and jumped to conclusions.

Wolfram|Alpha on the other hand consults its internal database which gets fed from various sources, but not by some web spider but by actual people. Therefore the chances that Wolfram|Alpha can't answer your question is much higher. But if it answers, the chances of it being correct are much bigger.

So here is my current personal  search strategy:

- If I search inside a special topic with a strong website, I use Google with the "site:" attribute to search that site.

- If I need exact facts I ask Wolfram|Alpha

- If I want to compare facts I'd consider Google Squared.

- If I want general reliable information I search Wikipedia (german and english, since the quality differs a lot but with different winners on different topics)

- If nothing of the above applies (or the search failed to return the answer I was looking for)  I rely on Google Search.

On a different note: Microsoft published a new search engine. Again. It seems to be a failure. Again. Except maybe if you are looking for images or videos