- From: James Clark <jjc@jclark.com>
- Date: Tue, 24 Jul 2012 10:31:41 +0700
- To: David Lee <David.Lee@marklogic.com>
- Cc: "public-microxml@w3.org" <public-microxml@w3.org>
On Tue, Jul 24, 2012 at 9:41 AM, David Lee <David.Lee@marklogic.com> wrote: > Good start! > I suggest discussing #7 ... > What does that mean exactly. Probably my ignorance more than anything. The wording I chose for #7 (MicroXML shall facilitiate the creation of documents that are simultaneously well-formed MicroXML and valid HTML5) is intentionally a bit vague. The ideal situation would be that MicroXML meets two requirements. 1. MicroXML can be used to represent any HTML5 DOM, and 2. HTML5 validity for MicroXML documents can be specified at the MicroXML data model level More formally, there is a set of constraints H on a MicroXML data model, such that 1. for any HTML5 DOM d, there is a string s such that - s is well-formed MicroXML - s is valid HTML5 - HTML5-parsing s produces d, and - the result of MicroXML-parsing s satisfies H and is isomorphic to d and 2. for any string s that is well-formed MicroXML, if MicroXML-parsing s produces a data-model d that satisfies H, then s is valid HTML5 and HTML5-parsing s produces a data model isomorphic to d. However, I suspect this ideal will be hard to achieve. One thing that makes it hard is elements like <script> in HTML5. The simplest data model for MicroXML would not distinguish between <script></script> and <script/> but HTML5 unfortunately requires an empty script element to be represented by the first and not the second. James
Received on Tuesday, 24 July 2012 03:32:30 UTC