What's really proposed at this point?
The kind of information that we'll have at first would be things like
author's names (possibly separated into first, last, etc.), date published,
created, maybe even last-modified, even though that's in the HTTP header,
key words...
The debate was between adding lots of new tags similar to <TITLE> or
adopting something like META that would be a more generalized scheme... or
using comments (ugh). I brought up semantic tagging, but that's overkill
at this point... although we should bear such needs in mind for the future.
How about a scheme that permits directed acyclical graphs?
META looks like this:
<META name="TITLE" value="Rotkaeppchen">
<META name="YEAR" value="1990">
With this sort of scheme, we can just teach our spider to look at the name,
see if it fits an attribute in the index being built and add the value if
it exists (or optionally, create the attribute and enter the value).
I just noticed that back in June when this discussion got hot, someone
suggested that www-html was a better list for discussing it. I'll
cross-post and let's do the follow-up there.
Nick