W3C home > Mailing lists > Public > public-html@w3.org > February 2010

Re: Updated DOCTYPE versioning change proposal (ISSUE-4)

From: (wrong string) Å„ski <kornel@geekhood.net>
Date: Wed, 17 Feb 2010 17:26:42 +0000
Cc: public-html@w3.org
Message-Id: <6291E035-39E0-41BB-98C7-DE1EB7220E03@geekhood.net>
To: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
On 17 Feb 2010, at 16:38, Leif Halvard Silli wrote:
>>> 
>>> It works fine in any browser and AFAIK it always did:
>>> 
>>> http://www.målform.no/html4-or-html5/index
>> 
>> By "work" I mean doing what internal subset was defined to do in 
>> SGML, rather than being misinterpreted or completely ignored at best.
> 
> So you insist on making points about SGML.

I don't know how can I make point about your use of esoteric SGML features in theoretically SGML-based document to (apparently) appease an outdated SGML-based validator, without referring to SGML.

Just to make it clear: I don't want HTML 5 to support SGML. In fact, I prefer the opposite, especially elimination of DTD, DOCTYPE (other than the minimal one necessary for standards mode) and SGML-based validators.

>> i.e. in text/html you are unable to use internal subset for its 
>> purpose. At best you can use it to entertain users validating your 
>> page in validators that are removed from reality.
> 
> Validation is very important. By adding a internal subset, I can make 
> the validation _more_ up to reality.

Yes, validation is very important. However use of SGML (HTML 4) validator on HTML 5 document should not be mistaken for validation of HTML 5.

If your validator is unable to parse HTML 5 documents properly, then you should be filing bugs against the validator, rather than expecting HTML 5 and every UA to change syntax to help you litter code with workarounds for some inadequate validator.

DTD-based validation is very poor and superficial, often mistaken for document conformance. IMHO should be discouraged and I quite like that in this case discourages use of hacks for DTD-based validators. I don't think anybody would accidentally use DOCTYPE with internal subset, so it won't cause proliferation of quirks mode, but will prevent proliferation of ugly hacks and misleading "validation".

-- 
regards, Kornel
Received on Wednesday, 17 February 2010 17:27:50 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:14 UTC