Re: algorithmic normative conformance requirements, design principles, etc.

On May 27, 2009, at 5:30 PM, Boris Zbarsky wrote:

> Larry Masinter wrote:
>> Specifying conservative authoring behavior at the same time  
>> encouraging
>> liberal reader behavior is the primary method by which standards  
>> writers
>> address those real concerns.
>
> This works well if the authors are sophisticated enough to be aware  
> of the specification.  In practice, that's not the case with HTML.
>
> That said, the current draft does in fact impose all sorts of  
> requirements on authors while defining what UAs should do if authors  
> violate those requirements.  So I'm not sure what we're actually  
> discussing here....

The spec has conformance requirements for markup, but generally not  
conformance requirements for use of scripting APIs. One reason for  
that is that we don't have good technology for validating a program's  
use of an API. In general, this kind of validation is equivalent to  
the halting problem and thus provably not computable. The best you  
could do is warn about violations at runtime, but that can't tell you  
about potential violations that can occur in different flows through  
the program.

Regards,
Maciej

Received on Thursday, 28 May 2009 00:36:58 UTC