- From: <noah_mendelsohn@us.ibm.com>
- Date: Thu, 9 Oct 2003 16:53:47 -0400
- To: xml-dist-app@w3.org
On Oct 1, at the request of the WG, I requested opening of an issue relating to canonical forms. The note is at [1], and we have indeed opened issue 441 [2] (thank you Yves!) Here I provide my personal opinion on this issue. The questions raised are: > On the call, there was debate as to whether it would or > would not be OK for the lexical form of an infoset item > to change during transmission. One opinion expressed > was: yes, no problem, c14n can be run on the sent or > received infosets in any case where you really care. > The converse opinion was: no, don't let it change, > because the SOAP Rec requires that all bindings (MTOM > or otherwise) transmit the Infoset with full fidelity. > So, we need to discuss whether to indicate that only > canonical forms can be optimized by MTOM. My answer is: * The SOAP Rec is unambiguous, to be a SOAP binding you must preserve the Infoset, which means you must preserve the characters even if in non-canonical form (especially since SOAP doesn't even know about types!) If it's an integer, you must preserve leading zeros, and if it's base64Binary you must preserve whitespace, etc. * The MTOM optimization seems incapable of distinguishing more than one lexical form for a value. From these two points I conclude that MTOM can optimize at most one lexical representation of each value. > If yes, we need to decide whether to adopt the canonical > form suggested in the errata to XML schema [1] or some > other form. Well, I can see the case for a form with no linebreaks at all, but on balance I think it's best to use whatever the Schema WG defines as canonical. So: "yes", use the canonical form as defined in the Schema erratum. Noah [1] http://lists.w3.org/Archives/Public/xml-dist-app/2003Oct/0001.html [2] http://www.w3.org/2000/xp/Group/xmlp-issues#x441 ------------------------------------------------------------------ Noah Mendelsohn Voice: 1-617-693-4036 IBM Corporation Fax: 1-617-693-8676 One Rogers Street Cambridge, MA 02142 ------------------------------------------------------------------
Received on Thursday, 9 October 2003 16:54:30 UTC