- From: Shane McCarron <shane@aptest.com>
- Date: Tue, 24 Jan 2006 20:05:43 -0600
- To: Anne van Kesteren <annevk@opera.com>
- CC: Shane McCarron <xhtml2-issues@hades.mn.aptest.com>, jim@jibbering.com, www-html-editor@w3.org, xhtml2-issues@aptest.com
(Speaking for myself, and not for the Working Group). Anne van Kesteren wrote: > > On Tue, 24 Jan 2006 17:01:00 +0100, Shane McCarron > <xhtml2-issues@hades.mn.aptest.com> wrote: > >> 2) we don't define error behavior, in particular in the case of >> validation >> errors. This is handled by XML. > > > At one point the HTML WG looks at current implementations and at > other points they simply miss the fact that implementations typically > do not have validating parsers. I disagree - the working group knows that most browsers do not validate. Nor is there a requirement that they do. But let's not have that argument - its circular and painful. Let's instead try this one: you cannot just say "add error conditions to XHTML2" - that's not a concrete proposal. A concrete proposal would be "Change the specification to read... something". Here's an example: Conforming user agents must refuse to process any document that claims to be an XHTML 2.0 document but does not conform to the requirements for a Conforming XHTML 2.0 Document. When such a document is encountered, the user agent must by default present an error message and not process or render any of the content from the document. I don't think that is the type of thing that anyone really wants, mind you. But you really should present a concrete proposal. You (and others) assert that it is incumbent upon a standard to define behavior in the presence of incorrect usage. As evidence for this, you seem to claim that without this it is impossible to have interoperability. That seems confused to me. Interoperability is about how portable applications/documents function on conforming implementations. The working group has attempted to define the requirements a portable application/document must adhere to (XHTML 2 Strictly Conforming Document) and the requirements on a conforming implementation (XHTML 2 User Agent Conformance). The drafts have all been very consistent in this; ever since the first drafts of XHTML 1.0 actually. A conforming document is one that is VALID and adheres to some other requirements. If a document does not adhere to those constraints, then it is not XHTML 2. It is an error for someone to assert that a document conforms when it does not adhere to those constraints. If it is not XHTML 2, we don't know what it is and we cannot define processing rules for it... Other than something draconian like the clauses above. On the implementation side, the drafts have required that conforming documents be processed. Neither you nor anyone else would want the recommendation to say, for example, that an implementation was not permitted to process documents that conform to other specifications. Nor would you, I suspect, want the specification to overly constrain your ability to provide value add beyond the basic conformance requirements (user interface enhancements, user customization, etc.) Moreover, if the working group were to place constraints such as the ones above on implementations, it would force conforming implementations to validate (how else would you know if a document was conforming?)... and no one wants that. The web is slow enough as it is. But you and others clearly have some point too subtle for me to have grasped so far. What am I missing here? -- Shane P. McCarron Phone: +1 763 786-8160 x120 Managing Director Fax: +1 763 786-8180 ApTest Minnesota Inet: shane@aptest.com
Received on Wednesday, 25 January 2006 02:10:21 UTC