W3C home > Mailing lists > Public > www-validator@w3.org > August 2005

Re: Relaxed - new HTML validation service based on RELAX NG + Schematron

From: Terje Bless <link@pobox.com>
Date: Sun, 28 Aug 2005 13:44:01 +0200
To: www-validator@w3.org
Message-ID: <D0202000d-1042-DC9CDB56BD2C43A2B66E9FB5D28836FA>

Henri Sivonen <hsivonen@iki.fi> wrote:

>I think dispatching on heuristics is a problem with the current W3C

How so? We deliberately keep heuristic behaviour to a minimum.

e.g. we do not dispatch based on an “xmlns” attribute when documents lack a
DOCTYPE Declaration; while it's been argued that this is “sooo much better” than
“that silly DOCTYPE Declaration” (I'm paraphrasing, obviously), it would be
ambigious and lacks normative definition.

While we do dispatch based on the document type name when better information is
not available, we do so supported by normative specification and with clearly
identified results.

That certain rogue groups have elected to deliberately sabotage established
international standards — by introducing an absolute dependance on heuristic
processing of web documents in conflict with this standard — can hardly be
blamed on the validator.

>It emphasizes validity as an internal property of a document
>instead of validation as a quality assurance tool.

Heuristic behaviour would be “Hmm. Let me see… What does this document look like
it's trying to be?” Validity _is_ an internal property of a document; it
wouldn't make much sense to treat it otherwise.

It's likely that there are other features that would be useful for a Quality
Assurance tool that would better be treated as external to the document. For
instance much of the WAI guidelines — or internal “house rules” and similar —
would be better specified as externally imposed constraints along the lines of
multiple checkboxes in the interface to enable checking against this or that
additional ruleset.

And for lack of normative schemas to check the things DTD-based validation
cannot catch, these types of checks are also likely candidates for external,
optional, conformance profiles against which a document can be tested.

For an excellent example of this see Nick Kew's Valet tools at, e.g.,

>I think it would be more useful to be able to as "Is this document SVG 1.2?"
>than being able to just ask "Is this document valid whatever it is?"

First, inserting the relevant DOCTYPE Declaration and submitting the resulting
document to the validator is asserting that it is “SVG 1.2” and asking the
validator to confirm or dispute this assertion. If you're not authoring in “SVG
1.2” but rather in “WHATever”, you have bigger issues than validation.

Second, if you for some reason absolutely need the format specification to be
externalized you can use the DOCTYPE override feature of the validator to ask
the question external from the document itself. Your “whatever” document can
easily be checked against any of the DTDs in the dropdown menu.

Yvonne: «Where were you last night?»
  Rick: «That's so long ago, I don't remember.»
Yvonne: «Will I see you tonight?»
  Rick: «I never make plans that far ahead.»
Received on Sunday, 28 August 2005 11:44:12 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:58:52 UTC