Action related to i043

I took an action to draft a 'be careful' proposal for resolving issue
i043 (comparing RefParams). I've posted a blog entry at[1] which
outlines the general problem. I think there are a couple of ways we
could go;

1.	As other people have suggested on another thread, remove section
2.3 entirely.

2.	Add text along the following lines;

	Comparison of Reference Parameters should be performed using
EXCC14N. However, EPR issuers and comparators need to be aware that
EXCC14N might yield false positives in cases where reference parameters
contain QNames (or other values that rely on in-scope namespace
declarations) in element or attribute content and the in-scope namespace
declarations for the EPRs being compared differ. Comparators that are
aware of the data type of specific reference parameters can populate the
InclusiveNamespaces PrefixList parameter to the EXCC14N algorithm with
utilized prefixes in order to get better results. 

3.	Add text per 2. above with the following additional sentence;

	EPR issuers are discouraged from issuing EPRs containing
reference parameters 
	that use QNames in element or attribute content.

I'm not yet sure how I feel about 1. I'm happy with 2. I'm a little
concerned that 3. goes too far.

Regards

Gudge

[1] http://pluralsight.com/blogs/mgudgin/archive/2005/01/31/5518.aspx

Received on Monday, 31 January 2005 17:22:37 UTC