<br><br><div class="gmail_quote">2009/4/4 John C Klensin <span dir="ltr"><<a href="mailto:klensin@jck.com">klensin@jck.com</a>></span><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
--On Wednesday, April 01, 2009 11:26 -0700 Lisa Dusseault<br>
<div class="im"><<a href="mailto:lisa.dusseault@gmail.com">lisa.dusseault@gmail.com</a>> wrote:</div>While I do not<br>
see any possibility of addressing the issues in the IDN context<br>
-- even if only because a better solution for one particular<br>
language would foul things up for others using the same script--<br>
the issues that Jefsey and his colleagues have raised about<br>
appropriate case matching for French are actually good examples<br>
of this: if users has learned that most mappings behave in a<br>
reasonable and predictable way, they will expects all<br>
mapping/matching operations to work the way they would<br>
predict... and be confused or irritated when they do not.</blockquote><div><br>John,<br>let me summarize our "French" position as our Chair puts it. Actually users expect IDNA2008 as a revision of IDNA2003 to work better for them than IDNA2003. In the case of the French script it does not. However, the support of Latin majuscules is an accepted Unicode approximation (to use upper-cases instead and non-user-decided casefolding) which conflicts with ASCII DNs case insensivity in the IDNA specifc case. This is why a negotiation is to occur some way, to decide if Roman uppercases are ASCII or Latin.<br>
<br>- If such a fundamental conflict was not noticed by this WG, while French is one of their nearest natural language, was disregarded by its members but you (in spite of us documenting it on-line), and banned by the Chair, only relevant language authorities can decide of them (like ASIWG, is currently doing for LAS).<br>
<br>- this negociation mecanism belongs to the missing presentation layer. The double look-up is an attempt to such a negociation. But its algorithm is wrong: if Ecole.fr and <a href="http://ecole.fr">ecole.fr</a> are registered which one will be picked for "<a href="http://ecole.fr">http://ecole.fr</a>"? Another solution could be the class usage you did propose. Another is the virtual classes JFC proposes with the "x.--".<br>
<br>Now, the Chair does not want this WG (1) to dicuss interoperabilty with our ML-DNS exploration (2) nor listen to users, we found that the only way to proceed is a complete review of the DN understanding which protects the DNS from confusion in usage and in management and makes it fully able to support the semantic addressing space as we approach it.<br>
<br>Rémy Renardin<br>----------------------<br></div></div>