Monday, 18 May 2009
Quite a bit of media frenzy is following the launch of Wolfram|Alpha – but it is not a web search engine. At first glance it looks like a web search engine and you could be forgiven for thinking it is a web search engine but try a typical search and you may be surprised by the results – once it figures out what you mean and that is one of aspects which makes it stand out.
Wolfram|Alpha is a computational knowledge engine – it bases its results on known facts and mathematical formula from its ever increasing information repository. Whilst working on an Open University project I got to use a product called flex a computational knowledge based system (but without the web based front end) and it was an interesting product and gives some insight into how this product might work behind the scenes. In flex, you put in all the information you know (create an information repository), create a set of relationship rules, put in some data driven procedures then use the knowledge specification language to ask questions and run simulations. Wolfram|Alpha takes this generalised idea but combines it with the power of cloud computing to provide a huge information repository.
When you enter a search, Wolfram|Alpha makes assumptions about the words and phrases you have entered before submitted it into its engine which considers its relationships and applies formula. This is important as a typical web search engine does not make these assumptions or try and relate too much to the phrase you have entered.
For example, take a simple question “What is a sausage?” – let us first ask Wolfram|Alpha.

It has taken the phrase I entered and assumed I am talking about a food product (some alternatives are offered), it has also made some assumptions about the type of sausage I might be interested in. It submits this information to its computational engine and has found various reference libraries so it can tell me some useful nutrition information.
Now, let us ask Google the same question – and what does it tell us? It tells us what one particular website thinks a sausage is and starts to list sites that might or might not talk about the word sausage (as a food product or not).
Metaphorically speaking – depending on what sort of information you are looking at you can either look around the library (Google Search) to see what books might be around or you can ask a librarian to point you to the right reference book (Wolfram|Alpha).
Try it out yourself at www.wolframalpha.com
Related
- The Next Generation Of People Search Reaches Out
- Which Desktop Search
- MOSS: Configuring Search
- MOSS: Fixing trouble with the Search webpart
- Do Not Overlook The Benefit Of A Helpdesk



I totally agree. I first heard about it from Newsy and then did a little digging- found out it launched today. I just gave it a couple questions and I definitely plan on using it for science/math work but it fills a niche that’s different from Google’s. They will both have their place as research tools in my toolbox.
As a side note, I think this could be one step closer to artificial intelligence. True, right now the connections and algorithms are input by people. But the fact that it makes assumptions, I think, is a step toward smarter computing.
I think along with Delicious and Google, Wolfram Alpha will be my 3rd “research tool” that I will be using.
It will be interesting to see what the future holds for this type of tool.
Jamie