I’m giving a lot of talks this fall, and they all started clustering around the idea of the “future of knowledge.”
Jeez. What a big topic. But putting aside the hubris of me talking about this for a moment – which is asking a lot – it’s a fun and challenging thing to take on. It requires synthesizing a batch of themes I’ve been talking about for a few years now.
I’m starting with a series of talks in September, two next week in Havana, then one at the UC Berkeley Center on Open Innovation, and then a series in Australia and Japan carrying into October. Below is the abstract for the Berkeley one – which is likely to serve as the basis for most of the talks I give the rest of the year. Figured I’d post it and get the phrasing out there early. As ever, comments welcome in the comment space or via email.
//////////////////////
We are seeing the transformation of knowledge from something that is primarily conveyed in paper formats into something else: a computable graph, in which the knowledge is written in formats that computers can understand and interconnect, based on the same technologies that underlie the internet and web. Paper technology simply contains expressions of ideas, but the very technology of paper makes integration of ideas very difficult, if not impossible.
Graphs allow ideas to “snap” together into larger and larger networks, which can in turn allow computers to help us interrogate the knowledge more effectively. There are competing technologies to achieve this, but the idea of “the paper” as the core container for knowledge is dying, and technology will be the killer. This transformation is happening first, like the transformation of documents to the Web, in the sciences.
The move to a computable graph as a knowledge storage technology holds enormous promise for open innovation. It can make innovations easier to discover and evaluate, serve as a network for low-cost transactions, provide new bases for credit and impact, and in general provide a technical infrastructure for the purposive inflow and outflow of ideas necessary for OI.
But this is “uncommon knowledge” – we’ve never dealt with knowledge this way, and it shows. There is a significant amount of legal and technical infrastructure failure to be addressed. And there’s a lot of barn raising to be done.
What I have noticed is the main problem with Open Data (or alternative forms of communicating scientific information) is giving the person who finds it sufficient context to understand exactly how it was obtained and what it represents.
It would be nice if there was a standard way to represent scientific information so that it were unambiguously understandable by machines. For the short term that’s probably not realistic. But it certainly is possible to mark up data in such a way that is understandable by humans. It can be as simple as linking to blog posts or wiki pages that elaborate on the purpose and details of the larger project.
I don’t disagree that new types of digital document promise new opportunities for “knowledge” development beyond our existing frameworks, which are all founded upon ancient “paper” based models. However, computable graphs are simply not going to do it without more formal foundational work. IOW, just expecting ad hoc informal graphs of the kind anticipated by the W3 (et al) to be at all useful is unreasonable and doomed to failure.
Steven, thanks – and for what it’s worth, we agree. That’s why my team at Science Commons is creating a deeply non-ad hoc graph and giving it away. See the Neurocommons web site for more, and subscribe to the list if you want to see what happens in the coming weeks.