Two
major themes emerge from the Jon Husband interview of Dave Snowden that
we introduced yesterday (available as a 30-minute mp3 file here).
One theme is the subversive nature of Web 2.0 as it impinges on
existing hierarchies – of control, of planning, of archiving, of
knowledge organization itself. We'll get to that next week.
A second theme deals with
the changing view of knowledge itself, as it reveals the immense added
value that derives from meta-information. The two themes work together,
Web 2.0 tools allowing people and knowledge to interact and grow
synergistically in ways that no hierarchy-based scheme of knowledge
management (KM) can compete with.
Jack Vinson seized on this second theme in his commentary on the interview:
“One theme that ran through the interview was the
importance of context in knowledge work. Context is usually removed
when you remove the human element, whether that is by archiving best
practices to a “database,” or by asking experts to “tell me what you
know” about a given topic, or assuming knowledge is a fixed thing as
opposed to an interconnected flow of many things. Context is one of my
favorite aspects of KM discussion as well: context gives you one
connection to the human element of knowledge that old-school KM tends
to miss.” – Dave Snowden interviewed by Jon Husband
One key concept considered in the interview
was the value of tags, and their usefulness in organizing knowledge. It
seems clear that tags are the shifting fulcrum around which transient
views of knowledge depend, and people change the tagging dynamically as
they add in the relationships between source and consumer. This brings
us back to the fluid relationships enabled by Web 2.0 tools, seeming to
spell the death of traditional KM hierarchy. As Luis Suarez noted:
”...over the last decade the most neglected word in
Knowledge Management has been … context. And from there onwards he
makes the connection on how Web 2.0 ‘makes the context in which you
receive and filter information and knowledge more critical, partly
because you can control that context by exchanging with people you
already know or trust.’” – The Impact of Web 2.0 on Knowledge Work and Knowledge Management by Dave Snowden and Jon Husband – Part II
As to whether computers can derive from a
body of information meta data as valuable as the supplemental tags
humans can supply, Snowden says no. As Jack Vinson notes:
“He’s seen that people tend to name and tag content
in different ways, depending on the request. When asked to provide ‘keywords associated with an item,’ people tend to provide words and
terms that aren’t in the item. I assume this is contrasted with ‘keywords in the item.’ Similarly, when people are asked to ‘name an
item,’ they provide words not in the item. Dave uses this as another
example of how people can provide contextual clues that cannot be done
by auto-summarizers and other tools.”
This challenges the assumption that software
will have to tag knowledge eventually because humans simply aren’t
going to do the work. Perhaps it depends on the value of the tagging
transaction between human and knowledge resource. Crowdsourcing initiatives in this area do show promising early results.
Our
guess is that software will watch humans and quickly learn to make some
very good guesses even about contexts not apparent from the literal
words in a body of information. Not much that involves human effort
will remain out of the hands of robots for long: humans are much too
smart for that.
