RE: Issue: Are structure queries needed?

I very much agree with your approach. In fact, I
believe we should extend it to the resourcetype property
as well.
(1) DASL is not the proper forum to attempt to
standardize queries on XML valued properties or even 
on structure valued properties.
(2) If DASL standardized query on XML valued properties,
implementations of that would have to make special
cases for each such WebDAV property (e.g., resourcetype
and lockdiscovery), because these are undoubtedly
not stored as XML values. That is very tacky at best.
(3) I agree with all of your stated reasons.

The test for resource type could be a query operator
that returned a Boolean value, and had one operand:

TestResourceTypeordinary resource/TestResourceType
TestResourceTypecollection/TestResourceType
TestResourceTypeindirect resource/TestResourceType

for example. (I'm not making a serious proposal about 
the spelling of any of the above identifiers. I would
accept nearly any reasonable spelling.)

Alan Babich
-----Original Message-----
From: Saveen Reddy (Exchange) [mailto:saveenr@Exchange.Microsoft.com]
Sent: July 22, 1998 2:31 PM
To: 'www-webdav-dasl@w3.org'
Subject: Issue: Are structure queries needed?




We have previously described that DASL cannot deal with XML structures
easily. For example, this problem arises with the lockdiscovery
property. This problem prevents us from solving a simple scenario: list
all the resources that are not collections. We then proposed several
solutions for this.
I would like to revisit the one I feel: (1) solves this scenario, (2)
does not raise the bar for server implementations, (3) does not attempt
standardize non-existent practice, and (4) is extremely localized (has
no impact to other properties).
I propose a property that simply evaluates to TRUE or FALSE depending on
the type of resource. 
Somewhat more formally .... 
The DAV:iscollection property evaluates to TRUE if the DAV:resourcetype
property contains the element DAV:collection, and FALSE otherwise. 
Thus, I propose we punt for the time being any more complicated
operators. Scenarios we have identified that tunnel into XML are, IMO,
not as funadmental as this one.
Comments? 
-Saveen

Received on Thursday, 23 July 1998 17:54:34 UTC