[Exist-open] Two further queries re near()

[Michael B. <mbn...@mb...>]

The tiny problem I mentioned in my last posting is that near() seems to
interpret  the third (optional) proximity parameter n as n+1. This means is
isn't possible to tell it to search for directly contiguous words only. e.g.
a query incorporating
near(.,"less commercial",1)
retrieves
 ====
More important than the well-known involvement of Italian financiers such as
the Riccardi and the Frescobaldi in the lending of money to successive
English feudal kings was a less obvious commercial influence.
====
despite the intervening "obvious" in the last line.

Or, to adopt the example in the docs, near(LINE, 'best friend') would match
"best of friends" as well as "best friends".


That apart, I am led to tempt fate by asking whether it wouldn't be possible
to allow regexes in the second parameter to near(). Since, as far as I can
understand it, near() [unlike matches()]  uses the full-text index like |=
and &=, would it be feasible to let it use the same regex facilities as
those operators? This would be hugely useful in my applications.

But I ain't complaining. I'm still jumping up and down with glee at what the
regexes and near() already let me do!

Michael
---------------------------------------------------------
Michael Beddow   http://www.mbeddow.net/
XML and the Humanities page:  http://xml.lexilog.org.uk/
The Anglo-Norman Dictionary http://anglo-norman.net/
---------------------------------------------------------