In these days of Google's domination of the Web, it is highly refreshing to see innovative new search technology emerging.
Perhaps surprisingly there are all sorts of other search engines around - from new ways of displaying results at searchme.com to easier ways to enter the question (that semantic approach) with natural language through a research project at MIT called START at start.csail.mit.edu. Part of this is down to the rise of the 'semantic Web', trying to make the Internet easier to use.
But one technology that has been promising much over the last few months has finally become public this week.
Wolfram Alpha is a new search engine developed by US-based British researcher Stephen Wolfram that has some high aims. Over the last few months, it has almost promised to revolutionise the Internet, but as it turns out, the revolution is pretty much the same way that Segway was going to revolutionise transport.
But Wolfram Alpha comes from a strong heritage, based on the ideas behind the Mathematica tool for handling highly complex calculations. Now Stephen Wolfram has moved that idea to searching the Web, dramatically improving the accuracy of the search and creating a 'computational search engine'.
So when you ask what the distance is to the Moon, instead of returning some pages with the distance on, it calculates the distance at this moment and gives you the answer. Similarly stock prices are presented almost in real time, culled from the Web.
However, if the area is not covered, there is no answer. So the size of the semiconductor market has no response, and in Alpha world a semiconductor is Dow Chemical's dielectric polymer, despite there being lots of socioeconomic data - that the UK's gross domestic product was £1.7trillion last year
Never one to let an opportunity to innovate pass by, Google fought back with Google Squared, with new ways to display search data. Now results can be displayed in a table very similar to Alpha.
The example that Google uses is a query for "small dog", which pulls up a table with photos, origins, weight and height of various breeds. This comes from analysing structures on the web which seemed to imply facts, then checking whether that 'fact' appears across other pages. This, as you can imagine, is a huge task but well suited to Google's army of distributed servers.
There are other changes coming through as well, from search options that provide timelines of the data, to a 'wonder wheel' that spins up a map of related topics. For example, the wonder wheel for the transistor pulls up links to integrated circuits, operational amplifiers, how it works, mosfets, diodes and resistors.
However, Wolfram Alpha comes up trumps in some areas - it will tell you that Google has 3.4bn page views a day from 440m visitors, with 57% searching for things and 23% using Google email. But in some areas it is a bit US-centric, using US census data. But if you want anything to do with numbers, then Alpha is the engine, and the simple search engine isn't so simple any more.
See also:
Nick Flaherty On: Venture capital vampires
Nick Flaherty has been covering technology since 1990 and is based in Bristol, where he co-founded the SiliconSouthWest network. During that time he has worked for most of the electronics magazines and newspapers in the UK and several in Europe and the US, covering all areas of the industry. He blogs at www.flaherty.co.uk.