Jim Hendler at NYLA and some futurist ramblings of my own…
Last week, at the NYLA 2008 Annual Conference I felt particularly privileged to have my mind blown by Jim Hendler of Rennselaer Polytechnic Institute. He talked to us about the development of the semantic web, the web science movement that is being touted as web 3.0. The semantic web is, in my opinion, going to change the way people interact with the world in a very big way. As librarians who care about things like community, public space, information, and context, we best be paying attention to this stuff.
So I believe the semantic web is going to permanently alter the way we all interact with the world. I’m still trying to get my head around it, but I’ll try to explain. The best way to start to understand the semantic web is to read the Wikipedia articles about RDF (Resource Description Framework) and ‘triplestores’ and to watch the Semantic Universe webcasts. Do not be intimidated, I am not a programmer and I do not write any code other than very basic HTML. When I read in Wikipedia that ‘triplestores’ are machine-readable subject-predicate-object statements, that suddenly made sense to me. RDF mimics your normal sentence structure but makes it something a computer can understand. Suddenly the English major can talk to the Math major: ie, your computer is closer to speaking and understanding your language. To quote the Wikipedia article, “a triplestore is optimized for the storage and retrieval of many short statements called triples, in the form of subject-predicate-object, like ‘Bob is 35’ or ‘Bob knows Fred’.” If you understand RDF and want to go deeper, look into Web Ontology Language (OWL).
Here’s how all of that applies to you right now. Look at Twine. I’d suggest that it is the first application of semantic web technology with potential for mass appeal. An explanation of Twine from the NY Times, a while ago:
Twine can scan almost any electronic document for the names of people, places, businesses and many other entities that its algorithms recognize. Then it does something unusual: it automatically tags or marks all of these items in orange and transfers them to an index on the right side of the screen. This index grows with every document you view, as the program adds subjects that it can recognize or infer from their context.
Context is what the semantic web is all about. Context is also what real people base decisions on. A honed understanding of the intricacies of context is what allows humans to make informed decisions rather than random selections as we bumble and boggle through our lives. As humans we don’t consciously ‘tag’ things or create logic statements in our minds, but as we move through time and events we unconsciously create reference points and associations and thus create context. The Twine ‘tag’ is modeled after this phenomenon.
When you break it down like that, sharing similar interests on Twine is nowhere near the most important long-term application of the semantic web.
The semantic web would have unique URIs for unique people/objects/items/entities, all of which are further defined by contextual metadata. So ultimately that’s a page, or a web location, for everything, and the metadata allows all of these pages to ‘speak’ to each other. This sets up the perfect framework for the forthcoming ‘physical internet’. Combine contextual linkage for unique people/objects/items/entities with real-world access points like Datamatrix or QR codes, GPS or RFID and there you have the best application of the semantic web. Hendler mentions in the Semantic Universe webcasts that the web 1.0 giant was Google, 2.0 was Facebook (or one of its competitors), and that 3.0 remains up for grabs. I want to suggest that the web 3.0 ‘killer app’ won’t be about search, it will be about people, prediction, and location, and it will take place in real space rather than on a traditional personal computer.
I suspect I’ll write more about all of this as my thoughts are further refined. If there are any taxonomists/ontologists out there reading this entry, I invite your commentary, corrections, and suggestions of how this developing technology might effect libraries and their services.
There are 11 Comments to "Jim Hendler at NYLA and some futurist ramblings of my own…"
You just blew my mind with this. Basically what you’re saying is that massive databases will be built and modified in real time based on tags generated by computer programs that will be able to ‘understand’ plain english syntax, right? Amazing.
From the world of a person who constantly has to come up with systems for cataloging video images (in documentary post-production), the only thing more amazing would be a program that does the same thing, but has the ability to recognize faces and locations, and to decipher speech. Web 4.0? Please?
I work with RDF, RDFS, OWL, and ontologies, and I think you’re right about how they will provide the platform for the next phase of the Web. Researchers and developers are working it from several different angles, from tedious and resource-heavy hand-coding to machine-generated tags that are then refined through use.
I suspect the real breakthrough will come when users are somehow encouraged, through whatever incentive, to label their content according to some generally accepts standard – in essence meeting machines halfway. And that incentive could take any form, from monetary to fun to simple ease of use (we all know what it takes to get us to do something). Furthermore, of course, there’s loads of legacy data that we’ll have to deal with, so automation will play a part if none of it is to be lost.
I also suspect a breakthrough will occur through mobile computing, which is becoming more the norm and demands small packets of richly described data. Depending on how rich that metadata is, our handhelds could conceivably one day become an network of devices that constantly share and flow – but don’t natively store – semantically coded data. For more on this: http://playfullibrarian.blogspot.com/2007/10/evolution-of-machines.html.
Sorry to hog comment space, but I have one more brief note. As for your point in your penultimate paragraph about people and location: bingo. All of our interactions on the Web – or as Tim Berners Lee is now calling it, the “social graph” – including search and social communication, will become highly place-based, local, and personalized. Or at least as personal as we’re willing to hand over information about ourselves. For a taste of this, check out what http://www.socialminder.com/ is doing with mapping our Gmail contacts with our LinkedIn contacts.
This entire concept, which is becoming more the norm with average users, but isn’t quite there yet, is even more reason why librarians and people considering the profession need to get away from the idea that the work is done from behind the desk. So much of how the new technologies function isn’t based on the technology itself, per se, but on how people *behave*.
Librarianship is in a position to take advantage of a unique opportunity to become a truly user-centered industry. This requires librarians to embrace the task of learning how to collaborate with users on a closer level than ever before, beyond the traditional notion of “outreach,” beyond the basic brand of “books,” to give users more control over the library environment — everything from the functionality of the web site to what types of spaces and programs the library offers — customizing the offerings to fit into the community’s local context. That can’t be done from behind a desk in a passive fashion, and it requires a paradigm shift on many levels.
While the technology is shiny, and can be fun to play with (for some, it’s a nightmare), it’s important to not lose sight of the human aspect here. It’s not just about what programs or sites can draw the lines connecting your network, what people do with that information is key.
I’d take a look at microformats (microformats.org) as a current application of the semantic web.
Terence wrote: “I suspect the real breakthrough will come when users are somehow encouraged, through whatever incentive, to label their content according to some generally accepts standard.”
I suspect the breakthrough will come when the computers can take users’ input and make it fit the standards.
I run a QR code website in the UK. The site allocates you your own unique trackable QR code which can be traced on a google map. Not exactly the web 3.0 killer app you mentioned but maybe a good start.
More info at :-
http://www.qrme.co.uk/qrme-dynamic-qr-codes.html
DerikB wrote: “I suspect the breakthrough will come when the computers can take users’ input and make it fit the standards.”
To some extent you’re right; it’ll have to be a hybrid approach, likely from several different groups of developers.
But Facebook and its competitors are already getting people to label their own content. Their users willingly enter their birthdays and addresses and favorite albums and books, and even connect themselves as “friends” to other users with birthdays and addresses and favorite albums and books. All of that is semantic data just aching to be exploited.
So perhaps the breakthrough is closer at hand than we think.
picked up this resource via the web4lib listserv this morning, might be good further reading for librarians interested in this topic.
http://semdl.corrib.org/Book/
“Libraries have always been an inspiration for the standards and
technologies developed by semantic web activities. However, except for
the Dublin Core specification, semantic web and social networking
technologies have not been widely adopted and further developed by
major digital library initiatives and projects. Yet semantic
technologies offer a new level of flexibility, interoperability, and
relationships for digital repositories.
Kruk and McDaniel present semantic web-related aspects of current
digital library activities, and introduce their functionality; they
show examples ranging from general architectural descriptions to
detailed usages of specific ontologies, and thus stimulate the
awareness of researchers, engineers, and potential users of those
technologies. Their presentation is completed by chapters on existing
prototype systems such as JeromeDL, BRICKS, and Greenstone, as well as
a look into the possible future of semantic digital libraries.
This book is aimed at researchers and graduate students in areas like
digital libraries, the semantic web, social networks, and information
retrieval. This audience will benefit from detailed descriptions of
both today’s possibilities and also the shortcomings of applying
semantic web technologies to large digital repositories of often
unstructured data.”
[...] get yourself thinking along these lines and to expand on my earlier post about the semantic web, check out Kevin Kelley of Wired Magazine’s presentation at the Web 2.0 [...]
Hey Nate, excellent post and nice job of breaking a very hefty topic down in plain English. I would offer one note– don’t forget that a very large piece of the Semantic Web puzzle lies in decentralized data availability and accessibility….the creation of a “data bus” that essentially operates as the web’s universal database (implying standards for data markup) accessible by any application or service on the Web (through a standard query language such as SparQL)….and that this data (including user identity data, their graph or relationship data, plus any other data that can be made publicly available) would provide near-intuitive context for users