Hi Günter,
my long-time project (kleist-digital.de) is nearly only based on documents (TEI-xml), so I'm working mostly with mixed content.
Thanks for the link to your project. I just remembered the (in)famous dash in Kleist’s Marquise (»Hier — traf er«), and it was interesting to look it up in one of the earliest editions of the text…
Are there any plans for future versions to implement the ignore option for full-text-search? It would help a lot.
From today's perspective, unfortunately, it has a rather low priority.
The implementation would probably raise various questions that would first need to be conceptually resolved. Right now, as you probably know, both the input and the query strings are simply atomized:
<a>A B</a> contains text { <b>B</b> } → 'A B' contains text 'B'
In order to ignore certain elements, we’d need to extend the atomization process to consider element borders. Things are even getting more complicated if data is indexed.
We usually solve this by creating additional index databases, which only contain the texts that need to be searched. The following XQuery expression shows how to create an 'index' database which all documents from 'data' without 'p' elements:
db:create( 'index', db:open('data') update delete node .//p, db:open('data')/db:path(.) )
Hope this helps, Christian