Dave_Raggett (dsr@hplb.hpl.hp.com) wrote:
[munch]
> I can think of one approach which would work efficiently.
> Lets assume that each section is defined something like:
> <SECTION>
> <H> this section's header </H>
> <SECTION>
> <H> a subsection header </H>
> </SECTION>
> </SECTION>
[munch]
Hm.. two things come to my mind here. First of all, I do agree
that having containers for sections would be a very useful thing.
If these sections were given attributes, like
<SECTION role="Abstract">
foo bar gronk ...
</SECTION>
a retrieval system would be able to handle requests like "give me
all documents that contain the term 'compression' in the Abstract".
This would be a further step towards intergrating more text database
functionality into the Web.
Secondly, I noticed that HTML+ defines a markup called <GROUP>,
which is intended to serve as a container for text sections. Am I
missing something, or is this basically the same as the <SECTION>
that Dave has proposed? Will group have some unique id that a
text database could refer to?
Cheers,
Chris
/*
* Christian Neuss % neuss@igd.fhg.de % ..in the humdrum
*/