Minutes for XML Core Nov 6 f2f morning (C14N 1.1)

Attendees
=========
Paul
Norm
Henry
Michael SMcQ
Konrad (IRC and Zakim)
Ed Simon (IRC and Zakim)
Frederick Hirsch
Thomas Roessler
Hal Lockhart

C14N 1.1
========
Frederick had sent out some suggested wording changes at
http://lists.w3.org/Archives/Public/public-xmlsec-maintwg/2007Nov/0000
http://lists.w3.org/Archives/Public/public-xmlsec-maintwg/2007Nov/0001
forwarded to the XML Core list by Paul at
http://lists.w3.org/Archives/Public/public-xml-core-wg/2007Nov/0017
http://lists.w3.org/Archives/Public/public-xml-core-wg/2007Nov/0026

Frederick explained his changes and took detailed notes.
We agreed on everything up to the point where we started
talking about removing Appendix A.

Konrad suggested a new algorithm at
http://lists.w3.org/Archives/Public/public-xml-core-wg/2007Oct/0024

Thomas suggested changing the third bullet in Frederick's
latest suggestion replacing "to prevent the creation of
an output that looks like a net path" to "combine multiple
consecutive slashes into a single slash."

Norm asks why we can't just say that the effect we want
is to minimize the value and leave how that is done up
to the implementor.  Henry suggests we could say that
the minimized version is the "shortest".  But we wonder
if that is enough guidance.

We note that, if we are going to replace appendix A with
a new algorithm, we will have to redo some implementation
feedback.  So some of us want just to delete Appendix A,
but then we need to be sure the prose says enough.

We thought perhaps we could add the examples at
http://www.w3.org/2007/xmlsec/interop/xmlsig-interop-doc/testcases.html#
XMLBASE_ANNEXA
to the spec to address the issue of dropping the 
appendix.  We decided to add these examples as an 
informative appendix.

We talked about saying to get the shortest that gives
the same result as the original.

A big discussion ensued.  Roy Fielding was missed.

We decided that we also need to add to the third bullet
the fact that we then add a trailing slash if the result
(before adding that slash) ends in "..".

We believe our suggested changes to the third bullet
point allows us to delete Appendix A.

ACTION to Frederick:  Update the redline version with
our latest decisions and resend to the groups.

ACTION to Glenn:  Produce a new editor's draft reflecting
the changes suggested in Frederick's updated redline.

Then we noticed a problem in the merging process (which
produces the input to the "Appendix A process"):
where 'Base' argument to join-uris ends with "..".
3986 merge will discard that ".." which is wrong.

ACTION to Thomas and Frederick:  Get implementors to
run this new test case and report the results.

End of time for this topic.

Received on Tuesday, 6 November 2007 15:47:29 UTC