WolframAlpha & Google: Like Apples & Oranges
The launch of WolframAlpha led many to consider the question of whether it will challenge Google‘s domination in search. Having now had a chance to try it out I believe everyone has been asking the wrong question.
My first few searches on WolframAlpha produced very poor results. That’s before I realized that it’s not a search engine. Instead it’s a computation knowledge engine. This means that it tries to compute the answers to questions rather than returning links where the answer may be found. It actually produces some pretty impressive answers as long as you ask it the right kind of question.
The best way to illustrate this difference is to describe an example. For a complete picture, I am also going to consider Wikipedia’s role in ‘information gathering’. Wikipedia, WolframAlpha and Google are complementary services that each provide valuable information, but in different ways.
Suppose you were interested in chemistry, and specifically the element Uranium. Here are some example searches that should illuminate the differences between the three.
In WolframAlpha you might pose the question “What’s the molecular weight of Uranium?“. Very quickly you will learn that the answer is 238.02891 (grams per mole).
In Google you could search for “Uranium Enrichment” and see links to the most relevant 10 of a total of 954,000 articles containing the keywords uranium and enrichment. These include government and educational information and much more.
In Wikipedia you might start by just looking for “Uranium“. There is a very detailed explanation of what Uranium is, and has links to related articles and entries.
You have three different searches with three very different but equally relevant types of result.
I predict that WolframAlpha will disappoint those who expect it to behave like (and/or better than) Google, but it will endure and establish it’s rightful place as a very relevant useful information gathering tool. I posted some comments on Twitter (@kfair) Friday night which expressed this disappointment. I now realize that reaction was premature and a little misguided.
My only complaint so far is that it feels a little incomplete, particularly with respect to the types and categories of questions that it cannot yet answer. The reason that it feels this way is that it is only in the early stages of development and, provided that it doesn’t flame out because of misplaced expectations, it should get better and better over time.
I’m am enthusiastic about it’s support for natural language searches and I believe that the introduction of WolframAlpha will prompt innovation at Google and elsewhere to further develop semantic interpretation of search queries.
A good place to learn more about WolframAlpha is this introductory screencast.












View Comments to “WolframAlpha & Google: Like Apples & Oranges”
#WolframAlpha & Google: Like Apples & Oranges « A Bit More Fair http://bit.ly/NRuNt
This comment was originally posted on Twitter
Agreed, WA & Google are very different as of now.
http://pranavsbrain.peshwe.com/2009/05/toying-around-with-wolfram-alpha.html
now you need to review BING!
Here’s my initial (uneducated) evaluation of Bing – Bad name. I will do a more thorough review after I get a chance to try it out.