SPARQLZ is a stealth technology project aimed to provide a graphical user interface for everyday users to assemble, edit, share and mash-up modular, persistent, real-time searches across the web of Linked Data. It's a side project by an independent team within a large data corporation, with dreams of spinning their work off as a startup.
It's a pretty hot idea: it's like Yahoo Pipes, for Linked Data - but easier to use and already populated with big sets of valuable information to mashup and parse. Linked Data is a growing field of datasets that are categorized with standardized markup, tied together and easily cross referencable by machines. The US and UK governments, news organizations, music data bases, social networks and other organizations are participating in the official W3C Linked Data community. Now SPARQLZ aims to make all that data easy to construct future-facing search queries for.
SPARQLZ is named after the SPARQL query language for structured data. The service also uses technologies like Yahoo Query Language and real-time push format PubSubHubbub. Search results can be delivered to SMS, Email, Webhooks, a Feed Reader or other SPARQLs.
How would I use a service like this? Just imagine stringing if/then statements together through the cloud of Linked Data (see below). As a technology publisher, I'd like to receive notification if and when SPARQLZ finds a photo (from Flickr or CORDIS) of a person from a known tech company (defined as the companies listed in Crunchbase), that's headquartered in a country with a lower than world-median GNP (per the CIA World Fact Book). Could SPARQLZ do that? In theory, it could do things like that all day long - and you and I could trade queries like that back and forth like legos to piece together whatever kind of stream we were looking for. It's a really exciting vision.
Linked Data is growing fast and has incredible potential as a development platform. Sir Tim Berners-Lee, widely credited as the key inventor of the World Wide Web, is now focused on the growth of Linked Data as the web's next step.
But in order to get beyond the borders of Wonky-land, Linked Data needs a good User Interface. For many people, the ability to set up dynamic queries, mix and match them and have them deliver alerts to various devices could scratch an itch that many of us didn't know we had.
Ok, so let's be honest. This is probably not going to be a mainstream phenomenon. But are there thousands, tens of thousands or maybe a larger number of people who could create value for themselves using a query construction and publishing model like this? Who could not or could not so easily create that kind of value on top of Linked Data today? I think there are. Maybe there are even millions of people who could capture some of the latent value in Linked Data thanks to a tool like this.
SPARQLZ is certainly intended as a way to democratize creation and use of something rich with value but previously too technically inaccessible for many people to use who might like to. In that it's of the same vein as Blogger and WordPress were to text publishing, as YouTube is to video publishing and Twitter and Facebook are to social activity feeds.
Might this be the project that makes Linked Data hacking something that far more of us can engage in? That would be great, but first the team will need to gather support and launch itself as a company. For the sake not just of this small team of data-loving dreamers deep inside a big corporation - but for the sake of all us data-loving dreamers who would love to use their tool, I hope they can do it.
The latest version of the Linking Open Data dataset cloud, as at July 2009, maintained by Richard Cyganiak and Anja Jentzsch.