Comment resolution: Updated Grouping and Formal docs

There have been two e-mail discussions going on in recent days 
concerning the POWDER docs. One in public and one on a team-only list. 
This e-mail, and the changes to which it refers, are designed to show 
that those comments have been taken on board and relevant changes made.

Semantic Extension in the Formal Semantics Doc
==============================================
Back In December, Eric Prud'hommeaux raised a 'minor nit' about the 
wording that introduced the Semantic Extension [1]. Stasinos has argued 
consistently that the original wording was correct but, given the weight 
of opinion, as accepted that the change can be made [2] to allow the 
document to move forward. I would like to record my personal thanks to 
Stasinos for this as I know he remains convinced that he is right (in a 
nutshell it comes down to whether you extend the RDF data layer or the 
application layer).

The revised Formal Semantics document is now checked in at [3].

Canonicalisation in the Grouping of Resources Doc
=================================================

This debate came about after POWDER was presented at a W3C team meeting 
last month. Thomas Roessler was concerned about the less than precise 
aspects of our canonicalisation steps. Dan and Eric joined in (the 
team-only thread starts at [4]).

The canonicalisation section is the section of this doc that has caused 
the most outside interest and has been shaped with the help of various 
folk from the internationalisation domain. /However/ the sections that 
were so altered were those concerning %decoding, Unicode, normalisation 
to Form C etc. What wasn't discussed with the i18n community was the 
short section at the end called 'Further steps.' It was /that/ section 
that caught Thomas's eye as it was here that we talked about context, 
knowing a URI was generated by a form submission and possible ambiguity 
of Punycode to IDN conversion.

I played around with this, using the canonicalisation routine I use in 
my processor and grouping tester [5, 6] as a guide (I made a separate 
little UI for it at [7]). This actually showed the issue well - both 
Andrea and I had been hiding behind the ambiguities ourselves and hadn't 
actually implemented any of the IDN/Punycode stuff ourselves. I /had/ 
tried to implement the canonicalisation steps one by one but actually 
had made a fist of it and done it in such a way that _created_ problems 
in some edge cases (like query strings that happened to include a + 
character). Once I stripped it down and used some standard methods it 
became a) a lot simpler; b) deterministic without ambiguity.

End result, I've made a couple of slight edits to the grouping doc and 
deleted the ambiguous section (and references to Punycode and IDNs). 
What's left is simpler, non-ambiguous... and works in both my and 
Andrea's processors.

Revised grouping doc is at [8].

Further comment notwithstanding, I am going to continue with final 
preparations for our PR transition request (which means working with 
Matt to make sure everything is in the right place, a little easier now 
that I can do some of this myself).

Cheers

Phil.

[1]http://lists.w3.org/Archives/Public/public-powderwg/2008Dec/0034.html
[2] http://lists.w3.org/Archives/Public/public-powderwg/2009Mar/0004.html
[3] http://www.w3.org/2007/powder/Group/powder-formal/20090309.html with 
diff-marked version at 
http://www.w3.org/2007/powder/Group/powder-formal/20090309-diff.html
(member-only links)

[4] http://lists.w3.org/Archives/Team/w3t-tech/2009JanMar/0057.html
[5] http://i-sieve.com/cgi-bin/processor.cgi
[6] http://philarcher.org/cgi-bin/powder-group.cgi
[7] http://philarcher.org/cgi-bin/canon.cgi
[8] http://www.w3.org/2007/powder/Group/powder-grouping/20090309.html 
with the diff-marked version at 
http://www.w3.org/2007/powder/Group/powder-grouping/20090309-diff.html

-- 

Phil Archer
http://philarcher.org/

i-sieve technologies                |      W3C Mobile Web Initiative
Making Sense of the Buzz            |      www.w3.org/Mobile

Received on Monday, 9 March 2009 10:47:13 UTC