Re: [Bug 5752] Parsing should be specified for future updates

Hi Mike,

On Jun 15, 2008, at 1:01 AM, Michael(tm) Smith wrote:

> Rob,
>
> I note that at this point Hixie has made the following technical
> assessment of your proposal:
>
>  These really aren't areas where we have any flexibility to be
>  honest. Parsing is an incredibly complex area and the
>  constraints within which we have to work are very, very tight.
>  The current parsing model was based on extensive research over
>  billions of documents and multiple independent implementations
>  and I don't see any way that we could change what you are asking
>  for.
>
> Are you disagreeing with that on grounds that Hixie is technically
> incorrect?

No, everything Hixie says there looks correct to me. Unfortunately it  
has nothing whatsoever to do with the bug as reported and to the  
extent that it does he undermines his own argument. Since the many  
independent implementations each do something different here with  
parsing, we already have to establish a parsing algorithm that will  
inherently disagree with one or more implementations. So I don't  
disagree with what Hixie says there it simply shows a remarkable lack  
of understanding about the topic (Henri’s really the parsing guy).

> If so, given that Hixie is actually the author the parsing
> algorithm, if you disagree with that technical assessment, I think
> it is necessary for you show that it's wrong, not for Hixie to
> show it's correct -- specifically, you need to demonstrate how we
> could change what you are asking for without breaking the parsing
> model.

Opera, Mozilla, WebKit and IE all parse differently. If we actually  
are going to specify parsing in an interoperable way and we want  
implementations to take the time to change their parsing algorithms  
(keep in mind that these implementations are already fairly satisfied  
with their content compatibility), then we should make sure the  
parsing algorithm is forward compatible.

I demonstrated earlier in the thread how this could be fixed  
(regarding parsing of unknown head elements and using the solidus for  
self-closing unknown elements). Perhaps Hixie could provide a sample  
list from the billions of documents he researched that would break if  
we provide a properly specified parsing algorithm. I would really like  
to see Hixie stay on topic on these bug requests. The responses simply  
look like roadblocks thrown up against everyone posting to bugzilla.

Take care,
Rob

Received on Saturday, 14 June 2008 23:24:12 UTC