Re: TAG Decision on Rescinding the request to the HTML WG to develop a polyglot guide

On 02/18/2013 02:55 PM, Jirka Kosek wrote:
> On 18.2.2013 20:34, Sam Ruby wrote:
>
>> I claim that the HTML parser in libxml2 is not sufficient to allow
>> interop on "unrestricted HTML syntax", and therefore it is *exactly* the
>> reason that there is value in a more "restricted HTML syntax".
>
> It's quite hard to imagine that average Joe (sorry to all Joes) Web
> developer who can't produce neither valid HTML nor well-formed XML will
> be able to produce polyglot markup which puts even more constraints on
> syntax.

There are two parts to the Robustness principle.

The HTML5 parsing algorithm is the best way to date to cope with 
arbitrary HTML pages composed by average Joe's.

The Polyglot document is the best way to date(*) to cope with arbitrary 
HTML parsers composed by average Joe's.

Also, don't confuse validity with interop.  At the moment, many of the 
validity checks imposed by HTML5 have little to nothing to do with interop.

> Clearly there are some scenarios where polyglot makes sense, but
> polyglot shouldn't be considered more then that.

The scenario is cases where the content you produce may be parsed by a 
parser other than a recent browser.  Parsers that may be imperfect.

>      Jirka

- Sam Ruby

(*) Note the words "to date".  Come up with something better, and I'll 
support that instead.  I suspect others will too.

Received on Monday, 18 February 2013 20:23:28 UTC