[whatwg] Re: Doctype FPI

[Ian Hickson]

On Tue, 17 Aug 2004, James Graham wrote:
>
> I agree with your points but not the conclusion that you draw. In order 
> to conclude that automated syntax checking tools (aka 'validators') are 
> harmful, you must believe that the average quality of documents produced 
> in the complete absence of a validator would be better than the average 
> quality of the documents produced when a limited degree of automatic 
> validation is possible.

Yeah, you may have a point there.


> Your concerns can be somewhat alleviated by providing a validation 
> service but:
> - Noting the limitations of the validator
> - Not providing a badge to indicate valid markup

And also making it check for things beyond what DTDs and Schemas can check 
for, yeah.

I guess we now have the founding principles for a WHATWG validator tool!


> I should just note that I believe that incorrect semantics are often 
> (perhaps even usually) more harmful than incorrect syntax.

Absolutely. Incorrect syntax can have defined error-handling behaviour. 
Incorrect semantics can only rarely be detected at all.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'