Following on the things that John Robb has been writing about knowledge management and web logs, I’ve really started
to use the term “knowledge logging” to describe what I do on the various
web sites I run. To be honest, I really don’t care about the web or weblogging
or Blogger or Manila or Conversant — what I care about are a) the things I
already know, b) the things I learn that I didn’t know, and c) the interconnections
and intersections among the things I know.
Now that I think about it, I’ve been doing knowledge logging my entire life
— but today the available tools make it easier than ever to collect and organize
knowledge. My current obssession is with metadata. Metadata is just data about
data. In my case, the data is the text of whatever web log entry I’ve made today,
and the obvious data are the title or subject of the article and the date.
Obviously dated entries are very common in web logs, but they are also extremely
limiting. For example, one of my personal obssessions is the possibility of
achieving immortality (or at lease a very long lifespan) through medical advances.
So when I read about a new medical advance on that front, I might post a few
paragraphs trumpeting this information. But suppose I want to sit down and look
at all of the immortality-related weblog entries? Most weblogs make that sort
of aggregation of related information extraordinarily difficult. For example,
I’ve visited a number of weblogs that have commentaries related to the 9/11
terrorist attacks, but none of the weblogs I visit have a page where
I can visit to see all of the posts they’ve made related to the 9/11 attacks.
Why? Because a) it’s far too time consuming to do manually if you’re posting
more than once every couple days and b) weblogging software generally doesn’t
make it easy to automate this process.
Conversant, the software I use
to manage my web sites, makes automating this process relatively painless. I
run a site devoted to the animal rights movement, AnimalRights.Net,
and this is a screen shot of the message entry screen that I see whenever I
add a new article. The Subject is pretty self-explanatory, as is the date the
entry is going to be marked as posted and the author (me). But the rest of the
fields really add a lot of power. I’ll go through them one by one.
The Label field just has a few choices — Article, Documentation, FAQ,
Private — that describe what type of post this is going to be. Ninety percent
of what I do with my weblog on this site are short articles of 400-500 words,
so they usually end up with an Article label.
The Description field is simply a short summary of the article. The
idea here is that if I’ve written a 750 word article, it’s probably better to
put a short 15-20 word summary on the weblog and then link to the larger article.
But this description also has a number of uses elsewhere — for example, I have
an archives page showing all the new articles by year, and inserting the text
of the description field there gives visitors a bit more information to decide
whether or not a given article is really what they’re looking for.
Organization, Topic, and People are fields I use to further categorize
articles. Is the article about PETA protesting some experiment designed to produce
a cure for AIDS? If so, I just click on PETA in the Organization field and AIDS
in the Topic field. From there it is a snap to set up a page dedicated to PETA
that lists only the stories that I’ve marked as being about it.
Using metadata in this way is far superior to simply running a basic
search on “People for the Ethical Treatment of Animals” and returning
those results. For one, such a method tends to return pages which aren’t always
relevant. Using that method, for example, would return several stories in which
the group is mentioned only briefly in passing. Second, it is extremely difficult
from my experience to use just a basic search to capture information about more
complex topics. For example, I write about a lot of things that generally fall
under the category of “genetic engineering” but I don’t necessarily
always use exactly that phrase. When I was still using just a basic brute force
search, I would either have to tailor what I was writing to fit the demands
of the searching method or simply abandon the idea of having a complete list
of articles related to genetic engineering. This method lets me put the categorization
where it belongs — in the metadata.
Isn’t this time consuming and a lot of work, though? Actually, no. Conversant’s
implementation of metada (it calls them “Custom Fields”) is as easy
to use as it is powerful. Setting up a field and defining the values would take
maybe 10-15 minutes, and adding a new value within a field (which I do regularly
as some new group or individual emerges as newsworthy) takes just a few seconds.
The only real obstacle to this system is adding metadata to entries that were
created before this system was in place. But since few people seem to be using
metadata today, this is a problem associated with its general use rather than
anything specific to Conversant (besides, there are immediate benefits to using
metadata even if you’re not going to immediately use it with pre-existing data).
The benefits, however, are well worth it — this system allows a single person
like me to easily target information to users. Many people who visit my animal
rights site, for example, are interested only in PETA or terrorist incidents.
Using metadata makes it easy to not only present all the PETA-related stories
on a single page, but also makes it easy to provide things like a PETA RSS feed.
