Factoetum Wants To Be Your Internet Servant
What’s key about Factoetum is that all of the items are pure data objects, so the meaning, context, and relationship of the information should be able to be quickly understood by both humans and computers. So a search for Paris, the city, for instance, won’t be confused with Paris, the Hilton heir, and the content returned will be relevant to the French capital. The aim of Factoetum is twofold, Dyson says: To change the way that content is created on the Internet so that results are relevant and actionable, and to change search. “The end user should be able to eventually take data items from Factoetum and reassemble them somewhere else into another kind of piece of content,” he says. “And these data objects should be programmable by the user. So you take Paris the city in France and add it to a page with the weather and they should know about one another and be dynamic.” On the search front, he envisions raising HTML's game. “HTML is unstructured and if we stay on that path you always need a company [aka Google, Microsoft] with a secret sauce algorithm to find the most relevant results, because HTML doesn’t give any data information about what is actually on the page,” he says. With more structured data you can get accurate results and do something with them. “One of the issues with most popular social networks like Facebook is the data is not structured so you can’t create applications” that mish and mash it, he says. Right now the information in Factoetum can come from users, from iTunes libraries, from imports of OPML files, and soon via an API that it’s developing. (Dyson says there are also talks underway with potential business partners about importing product information into the site in real-time, as well.) All items are data and all items can be linked to other items, and these can be related by favoriting to members that are also data items, he says. Because the items are data objects, it’s easier to create algorithms about what can be done with them, because as a developer you know what the expectations are around particular kinds of items, he says. A Switchboard feature in the service gives users a choice of what they want to do with results – say, for instance, it’s a musical artist. A logical thing you might want to do for that is find related events -- for example, to quickly discover where the band is playing and when and act on that information. “If you type that into Google or Yahoo you just get a list of results, but you couldn’t take action on those results,” Dyson says. “But because it’s data and we know what’s coming, there is the appearance of artificial intelligence. And once you get to that point you can make all kinds of cool algorithms to make what people want to do easier.” The Switchboard part of the application suite that’s being built will be opened up to developers so that they can build other capabilities on top of Factoetum’s search and data services that members can then rent. “So instead of spending hundred thousands of dollars on R&D as Google and Microsoft do, and having a mono-directional idea of what the user wants, what we say is the user should be able to pick from a broad base of applications and services, and we want developers to help us and help themselves too,” says Dyson. For instance, applications could be created that show results for a search on Bordeaux (the city in France) all the restaurants that have seats open on a particular Friday night, leveraging Factoetum data objects and the data on the Internet that is available via third party APIs. He says most Switchboard apps can be created by developers in 24 hours. A micropayments system inside the service would let a developer receive payment each time the application is used. “I’m using the word rent because I don’t think you should pay for the whole price of an application if you only use it one time,” Dyson says. Maybe you could rent it one year, six months, per use, for example, and the onus is on the developer to make sure it works really well so that users do decide to rent it again after the first try. Dyson knows there are still some questions around the real-world validity of micropayments models, but says he’s willing to experiment, helped by the fact that “the overhead for the site is really low,” he says. The revenue model for Factoetum could break down into transaction fees from application developers, affiliate-based purchase programs, and also opportunities to provide licenses to enterprises that see an internal use for the service. Says Dyson, “I’m not going to leave anything on the table.” Email This Post |
The Voice of Semantic Web Business
|
|||||||