Re: [xsl] identify sections in an xhtml document

["Joris Gillis" <roac@xxxxxxxxxx>]

Tempore 01:28:30, die 02/11/2005 AD, hinc in  
xsl-list@xxxxxxxxxxxxxxxxxxxxxx scripsit Dean Maslic  
<dean.maslic@xxxxxxxxx>:

> Im thinking in a generic way, with any site.
> Some ideas I had were eg. calculate total num of nodes, then go
> through block level nodes (div, table,tr, ol etc) and calculate a
> ratio between their number of nodes vs. total number of nodes. If the
> numbers are roughly the same (say > 0.9), don't label, go to the child
> nodes and apply the same. If they are different, look for collections
> of of links (eg. count(descendant::html:a) > 5) or size of text nodes
> etc.
> Im sure there would be a way to do it for a generic 'standard' site
> (ie.page that contains a Top link-bar, left/right sidebar, and some
> text/image content)
Hi,

The algorithms you have in mind can be applied with XSLT, but I doubt  
they'll ever result in something usable.
It will only work with structured and well-designed -consistency - sites  
that are in XHTML. But such sites typically already have decent structure  
and/or well chosen class attributes from which you can easily derive it.

I don't think it will ever work with a "standard" website, which tends to  
equal a messy and bloated tag soup.


regards,
--
Joris Gillis (http://www.ticalc.org/cgi-bin/acct-view.cgi?userid=38041)
Veni, vidi, wiki (http://www.wikipedia.org)