Nova Spivak

Nova is a semantic Web pioneer and publisher of Twine.

Library 3.0--the Semantic Web.
Twine.com is based on the Semantic Web, create public or private collections. Combine the wisdom of crowds with automated annotations.

The social graph just connects people.
The semantic graph connects everything.
Nodes and links.

Everything is becoming metadata. A higher resolution Web. (Like megapixels on cameras--a more granular interpretation.)

In the semantic web we're putting the metadata into the data. Transforming the Web from a file server to a database. (But hidden from other applications.) RDF and OWL are the languages.

A programmer has to understand the API, do some work...but on the semantic web our data can talk to each other easily. The Web for machines. Software will be able to understand the Web.

Smart Data. Get humans out of the loop--it's not going to scale if humans have to do everything. The Smarts move into the data itself rather than being hard-coded into the software.

(The Evolution of the Web: Connections between information and Connections between people.)
*PC Era--e-mail. (Front end decade)
*Web 1.0--early back-end issues. (Back-end decade)
*Web 2.0--making the Web more friendly, more interactive with rich media (Front end decade)
*Web 3.0--another shift. (Back end decade)

Web OS will be like middleware. When the Web starts functioning like an operating system.
*Web 4.0--the Intelligent Web. Ubiquitous personal assistants. The notion of the Web spilling out over everything and making everything smarter.

5 main approaches to semantics:
Tagging
Statistics
Linguistics
Semantic Web
Artificial Intelligence

Some approaches make the data smarter, others make the software smarter.

The future of Information Retrieval...the tools make it easy to make MORE information. Leveraging the social Web: but as we move towards Web 3.0...we can use machines.

All human knowledge, of all forms, will be represented on the Web in 20 or 30 years.
How do you add Semantics?
(At Twine we use all of them.)
Manual Tagging--Hard/precise, tagged by experts. Easy/fuzzy, tagged by nonexperts.
Automatic--Hard/precise, database integration

Libraries that learn:
The library brain, fires in slow motion. (Librarians have to have ideas, then put them into place, then people notice, etc. Soon they can self organize.
*Digitize everything and put it all online.
*Automatically self-assemble and self-organize.
*Build personalized experiences for users.
*Detect and adapt to changing needs and usage

As users do this, recommendations and filtering become really important.