Wolfram Alpha

Tuesday, 19 May, 2009

Wolfram’s Alpha has made quite a big splash on the search engine scene and looks likely to redefine the whole search business. As with Google, the interface is minimalist belying what’s going on in the background and it doesn’t brand itself as a search engine, but rather a “computational knowledge engine.” And whtat is really the lead to take here; in the same way that Ask Jeeves tries to understand your question so too does Alpha. Except that instead of leading you to pages with the answer, it tries to answer the question itself using (not surprisingly!) some clever mathematics and a large array of source data.

So type in life expectancy and you will get back a page rich in detail showing you life expectancy around the world, a histogram and rank, all of which can be download as a PDF that is generated on the fly. This is very impressive and extends to a whole variety of subject areas all of which are worth exploring. So, for example, you can directly input equations for solving or graphing. On the geography side, you can geocode points, get elevations or produce maps.

It sounds like a panacea which it isn’t. For starters it’s quite slow at the moment, probably because of the publicity, however this has had the effect of repeatably killing (completely!) Firefox. Also, the results are heavily US centric with far less data available at national and sub-national levels outside the US. This is to be expected and something that will no doubt be worked upon. Each results page also has a “source” section; if you creating a “knowledge mashup” the data sources that are being mashed need to be far more explicitly referred to. To be fair there is a note stating that this can be requested, however if I ask for elevation kingston upon thames it would be nice to know where that information has come from. The effect upon derived/computed products can obviously be profound. This is a hugely rich site and I know will require a fair bit of time to learn how to fully utilise it. V2 will be interesting!

GPS Failure?

Monday, 18 May, 2009

Tidbits are reporting on the potential for a degradation in GPS accuracy over the next 2 years. The basis for the article is a recent Government Accountability Office report on the US Air Force satellite replacement program for GPS which is behind schedule and over budget (nothing new there then for the military!). However a risk analysis is highlighting the likelihood for the constellation to drop below 24 satellites potentially reducing accuracy. And, if over runs continue, the risk of a failure of the system. Not much to do but watch and wait, but clearly it’s a highlighted risk.

OS Reviews Free Data Model

Friday, 15 May, 2009

Charles Arthur has an interesting article over at the Free Our Data campaign, briefly summarising an invitee only presentation by the OS Chairman outlining their new “hybrid” strategy. Charles focuses upon the claims that the OS has reviewed the free data model, finding it non-viable. Quite rightly, Charles notes that this forms an important document within the debate on free data and should be made generally available. Assuming, of course, that it does exist in any formal sense. Well worth reading and pondering on.

IKONOS Data licensed by ESA

Wednesday, 6 May, 2009

Interesting news over at Mundogeo announcing the licensing of IKONOS imagery by ESA. The interesting aspect is, to quote the article:

“Ikonos as a Third Party Mission is available to Esa and Esa-approved projects for scientific research, applications development, or research and development in preparation for operational use in future, the so-called Category-1 use.”

Which means that an ESA approved project can have access to IKONOS imagery. Potentially a very useful source of data.

What’s your DOI?

Monday, 4 May, 2009

I would hope everyone in academic knows what a DOI (or digital object identifier) is. And to quote Wikipedia, it’s a “permanent digital identifier given to an object.” This is actually a very wide remit dealing with a vast number of “name spaces”, but most academics associate it with the funny string of numbers at the end of a reference in a journal article, thus:

Skidmore AK. 1989. A comparison of techniques for calculating gradient and aspect from a gridded digital elevation model. International Journal of Geographical Information Systems 3(4): 323-334. doi:10.1080/02693798908941519

This is a persistent identifier, so simply prefix it with http://dx.doi.org thus:

http://dx.doi.org/10.1080/02693798908941519

to be taken to a landing page for the article. All very clever stuff and it makes data discovery (and recovery) much easier. Of course it means every article needs a DOI and the publishers need to assign them. As publishing is just one area of DOIs, there is a registrar that deals with this on behalf of the DOI Foundation called CrossRef. Publishers enter article level metadata, including the DOI, set up a landing page and… voila, it all works!

Of course to make the most of the system it helps if all authors, when they write their articles, attach the DOI to all the references as well, however it can be a pain to trawl through all the papers and find this. Well CrossRef have a variety of DOI tools available, including one for a “reverse lookup”. That is, feed in the article details and retrieve the DOI (rather than the usual other way round). And they use some clever processing from eXtyles to achieve this through their appropriately named Simple Text Query. Copy all your references and paste them (in one go) in to the text box and press “Submit”. For every article it uses some clever algorithms to work out exactly which paper it is and then matches it to the DOI database returning the full details to you. An excellent service that just works faultlessly; invaluable research tool.

Ciao Bella, Bloomsbury

Friday, 1 May, 2009

I had a very pleasant meal at Ciao Bella, a wonderful “little” Italian restaurant in the heart of Bloomsbury and at the north end of Lamb Conduit Street. We went early evening when it was relatively quiet, but apparently it gets packed out and needs to be booked in advance. The food was excellent and, amazingly, you can get a two course meal for 2 for under 30 (without drinks). Highly recommended.