Re: ISSUE-4 - versioning/DOCTYPEs

Henri Sivonen, Thu, 20 May 2010 05:57:41 -0700 (PDT):
> "Leif Halvard Silli" <xn--mlform-iua@målform.no> wrote:
> 
>>> What KompoZer and NVu do is not a use case.
>> 
>> The use case is "an in-document and editor independent signal to
>> create polyglot HTML5".
> 
> I meant what [ARE] you are trying to accomplish that can be helped by 
> having an in-band indicator for requesting polyglot output from an 
> editor.

You seem to ask what I want to accomplish by requesting polyglot 
syntax. Is this just another way of asking for the purpose of using 
polyglot syntax in the first place? I prefer read it as if you asked 
why, given the existence of a polyglot syntax, there also is need for a 
indicator for this syntax.

The purpose of something that signifies a certain encoding, is of 
course standardization around that encoding. By inserting the polyglot 
syntax indicator, one indicates what coding flavor to use. It is very 
irritating if you seek to follow a certain standard, and then suddenly, 
gets your standardized documents converted into another standard, by a 
tool which didn't know that you tried to follow another standard than 
the default one - because you were only using a indicator that was only 
a private indicator. If there is a standardixed polyglot syntax 
indicator, then I can pick the tools which respect that doctype and 
tell my contacts and clients to look for the same tools. Instead of 
defining my own private standard, with my own private indicator which 
again requires my own private or proprietary tools.

>> I will answer a question that is relevant to my position: "How does a
>> <!DOCTYPE html 'polyglot-signal'> enable validation more quickly?" 
>> Answer: Anyone that is able to create a DTD can create a validation 
>> service today. A DTD is a schema. For e.g. Validator.nu, we must wait
>> for your excellent work. (Yes, DTD validation is in some ways more 
>> superficial than Validator.nu.)
> 
> That's not so. With Validator.nu, you can use your custom RELAX NG 
> and Schematron schemas. The parts of Validator.nu whose behavior 
> can't be customized via RELAX NG are also uncustomizably via DTDs in 
> DTD-based validators. 

I stand corrected, then. 
-- 
leif halvard silli

Received on Friday, 21 May 2010 00:06:15 UTC