Hi Justin,
On May 9, 2008, at 2:37 PM, Sam Kuper wrote:
> 2008/5/9 Justin James <j_james@mindspring.com>:
> Anything that cannot be machine-checked should not be in the spec.
>
> Justin,
>
> This is bonkers. Large parts of the spec can't be machine-checked.
> How can a machine detect whether text in a <p> element is being used
> as a heading, for instance? I could go on listing cases, but I won't.
>
> I think you may be confusing a schema (something for computers to
> use to assess validity) with a specification (something for humans
> to use as guidance when writing code or markup for computers).
Another way to put this might be to say: for machine checkable only
merely requires SGML and XML. We would need no HTML specification.
Most of HTML5 (perhaps all of HTML traditionally), specified the
mostly the non-machine checkable parts (except for the validity in the
XML sense of separating validity from well-formedness).
Take care,
Rob