element proximity

["Ray Denenberg, Library of Congress" <rden@loc.gov>]

Looking at the LOM example:

dc.creator = sanderson and the dc.date =2006 are contained within the same
contribute container.

I have some questions.

First question, what does it mean?
Searching for instances of element <contribute>,  with both subelements
 <dc.creator> sanderson </dc.creator>    and
<dc.date>2006</dc.date>


Second question:
Is this something that XQuery does well?


Third:
In Kerry's paper, a suggested CQL solution using proximity is:
dc.creator = sanderson prox/unit=element/distance=1 dc.date=2006.

element/distance=1 intuitively says that they are consecutive elements
(despite the eloquent discussion to the contrary in the paper).

So, I think I would rather use unit=parentElement/distance=0   (they occur
in the same parent element) and I would think with this approach 0 is the
only meaningful value for distance, so it can be omitted.

But the more troubling problem I see is: how does this nail it down to a
<contribute> element? In other words this would find any element in which
these two occur, not necessarily  <contribute> elements.

Let's say we are (hypothetically) going to define an "element" context set
for this, so:
PROX/element.unit=parentElement

Then perhaps define another prox modifier in the element context set:
parentElementName, so:
PROX/element.unit=parentElement/element.parentElementName=contribute

or just
PROX/element.parentElementName=contribute

Thus, you may be looking for two elements together, regardless of the parent
element name, or you may be looking for them together within a specific
element.   In the first case, use /element.parentElement and in the second
/element.parentElementName.


For discussion Monday.

--Ray