Re: Client-side highlighting; tag proposal

John Franks (john@math.nwu.edu)
Fri, 10 Mar 1995 09:30:12 -0600 (CST)


From: John Franks <john@math.nwu.edu>
Message-Id: <199503101530.JAA03594@hopf.math.nwu.edu>
Subject: Re: Client-side highlighting; tag proposal
To: www-html@www10.w3.org
Date: Fri, 10 Mar 1995 09:30:12 -0600 (CST)
In-Reply-To: <ab8564ae04021004a16a@[192.187.143.12]> from "Nick Arnett" at Mar 9, 95 10:43:10 pm



A feature quite similar to what is being discussed here has been
available for some time in the WN server.  Matching items from full
text searches are enclosed in named anchors.  This has the effect of
highlighting them, but also allows the browser to bring up the
document with positioned at the match.  Thus the response to the
search is a list of documents with sublists of lines containing a
match with the search term highlighted.  Selecting the highlighted
term takes you to the relevant document postioned at the match.
Selecting the highlighted search term in the document takes you to
the beginning of the document.

You can try this with a demo which searches the 39 files of the 
HTML 2.0 specification at

     <URL:http://hopf.math.nwu.edu/html2.0/dosearch.html>

By the way, tagging a search term match is not a simple matter
in HTML.  Very often the match is in an anchor which does not allow
an additional tag.  And simply deciding whether or not a match is
in an anchor requires parsing the whole document in principal (but
a few neighboring lines seems to work fine in practice).  Then there
is the question of what to do if the match is in a place which 
allows no tag.


-- 

John Franks 	Dept of Math. Northwestern University
		john@math.nwu.edu