Mohit,

Sorry to take so long to reply - I'm am just now finally processing my list of XRI Resolution edits.

See my answers marked ### inline below.

=Drummond

_____________________________________________ From: Mohit Raj [mailto:mr...@amsoft.net] Sent: Friday, January 20, 2006 5:02 AM To: xr...@lists.oasis-open.org Subject: [xri] XRI Resolution 2.0 Draft 10 comments

Hi all,

My questions regarding the newly submitted draft are below:

1. Section 4.2.2 - Line 511 states that "Parameters specified in this manner SHOULD take precedence over the same parameters provided via any other interface". Of the 3 optional resolution parameters, if only 1 is provided in the HXRI, the values of the rest should be considered as null or other interfaces (such as HTTP header) should be considered?

### Yes. The rule should be: if the parameter is supplied in the QXRI, then it takes precedent over the parameter supplied via any other interface. This includes null parameters, i.e., if the QXRI includes a parameter that explicitly sets a null value, then the parameter should be set to null. However if a parameter is completely absent from the QXRI, but is supplied via another interface, the parameter value should be accepted from the other interface. ###

2. Section 5.6 - Flowchart suggests that the value of xrd:XRD/xrd:Service/ xrd:Path should be equal to either Path OR "(" Path ")", in order for the service to be selected. Whereas the normative rules for the algorithm, pt. 4 says that the selection of a service should be *repeated* with the value of path = "(" Path ")". So, is there an EQUALITY or a PRIORITY relation between the path values distinguished by parenthesis?

### Great point. It should be an EQUALITY relation. As a result, if you had the following two Service elements...

<Service> ... <Path>+contact</Path> ... </Service>

<Service> ... <Path>(+contact)</Path> ... </Service>

...then if the QXRI path value was "+contact", it would select BOTH of the above services as equally matching this path.

=Drummond

<<attachment: winmail.dat>>