Silk is a web based tool currently in private beta for adding and extracting more structured information from the sites which matter to you. Once a site has been imported into the Silk application, it allows complex queries and filters on that data to be performed. Think of it as something of an additional semantic layer built on top of your pages that lets you ask more detailed questions of the information contained within. Silk effectively lets you identify segments of data in your site, and makes them universally searchable. For example, in the case of ‘The Simpsons’ (one of the sample data sets that I played with in the beta version) – I was able to run a query which showed me:
“How many times did Lisa appear in Episode 1 of the Simpsons?”
This is the sort of query that Google fails miserably at. However (provided there is a data set available) Silk excels.
Silk plan on leveraging the content provided by others to grow their semantic database, and also provide hosted copies of their technology to publishers. TheNextWeb for example have launched the first public dataset on the web today, which gives you a feel for its application.
As we try to get a grip on more interesting ways in which semantic data can be utilised Silk have provided a platform to get a headstart. If you currently present a database of any kind in plain HTML that you may find their offering an interesting way to more easily query that information, and allow your visitors to do the same. Provided they get adoption, and a healthy selection of datasets, Silk may just be the semantic query engine the web needs.