Tuesday, May 24, 2005
Notational Standards
discussion à
The next generation knowledge tools
John Sowa said: [87]
But the point I derive from these and many
other observations is "Who needs tags?" A tag is nothing more nor less than somebody's attempt to
classify a text according to some preconceived notion, ...
I'd argue that a tag is less than that, in a really important way.
A tag is a label. It may also be a classification, but it's often not, as with tags like 'to_read' or 'for_drew', or even 'clayshirky', which tags are reminders, routing instructions, and source info, respectively.
Sowa also remarked:
As a result, I think the idea that a tag "[...]
has a good chance of being (a) wrong, (b) irrelevant, or (c) misleading"
is (tautologically) untrue.
If it's my tag, then it
can't be wrong, so long as I say it's not wrong.
Where it _can_ be wrong is
in someone else's context. For instance, a number of people tagging 'Ontology
Is Overrated' have tagged it 'taxonomy', which is incorrect from my point of
view, but not from theirs.
Since all the tag data includes user sources, I'm free to use only my own tags, to whitelist a group of users I trust, to blacklist a group I don't trust, and so on.
Sowa also remarked:
You can derive an enormous amount of information from
the text of the web pages, and the tags just get in the way.
I just put up a brief discussion of the tags forming around that piece, ( http://tagsonomy.com/index.php/dynamic-growth-of-tag-clouds/ ) including noting that a number of the tags -- e.g. IA (for information architecture), metadata, Semantic_Web, and so on -- are perfectly sensible assessments of the larger context of the piece, even though none of the tagged words appear in the text itself.
So the tags that refer to user context, as well as those that are 'notes to self' such as to_read, are not extractable from the pages themselves.
Clay Shirky