> So, again, what tags should we use? What do we want it to
>look like when we convert the info from your_favourite_directory_engine
>to HTML/SGML?
We need some resolution on the entire tagging/META issue fairly soon. Our
engine creates indexes that include document attributes, which typically
are things like author, date created, etc. Our spider will extract some of
them from the HTTP header, but it's also going to look in the HTML header.
We are going to have to commit to some baseline specification very, very
soon. One of our short-term goals, for example, is to build a searchable
"virtual library" of documents on the Web that deal with the Web,
publishing and libraries. We'd like to be able to tell the owners of their
pages how to put tags into their headers so that our spider (and others)
can gather the attributes. And we'll be doing that with our own documents,
too.
I wouldn't want to do anything that would be un-SGML-ish, but I'd also like
to build a basic structure that won't rule out semantic and relational data
models and cross-industry standards such as MARC.
The momentum will take a bit of a leap on Tuesday, I suspect, when we'll be
making an announcement with Mosaic Communications.
Nick