I think we are moving a touch away from the point here. The original proposal
was for a quick and easy search method. Agreed that it should probably
expandable in the future, but I think that there are 2 issues here.
1) A quick search. This is probably similar to the <A NAME="..."></A>
thing that we have all been using to jump to points. If I recall the original
suggestion it was that we could specify a *SINGLE* keyword and use it
to do a textstream search of the document to jump to a point. Actually
thinking about this last night, I would perhaps propose a slight change to
this. Something like:
http://a/doc.html#text=string A simple text search
Then we could expand to say:
http://a/doc.html#image=image.gif Jump to a picture
http://a/doc.html#url=//host/file Jump to a given link
http://a/doc.html#div=number I always wondered what to
use <DIV n> for :)
which we could then use to bring us back a full circle to (you guessed it)
http://a/doc.html#name=tag Which is the same as:
http://a/doc.html#tag
2) The full document search (including child documents). I leave this up to
all to come up with there own scheme. Remember that the _original_ HTTP
spec mentioned being able to extract areas from within images too. I had
a few ideas on that but trying to work them through came up with hurdles at
every angle.
But this is a very interesting topic, and I dont think that however it turns
everyone will be happy.
Paul