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

On Thu, May 28, 2009 at 12:06 PM, Larry Masinter <masinter@adobe.com> 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.
>

For HTML, specifying conservative authoring behaviour will have almost no
effect on actual authoring behaviour. This is another unfortunate fact we
can do nothing about. This is different from most other computer-related
specifications: someone assembling and sending TCP packets can be expected
to follow advice given in the spec, but someone sending HTML cannot. So
rules of thumb from other spec domains do not necessarily apply here.
(Perhaps this is one reason why HTML5 has had more emphasis on
implementation requirements than author requirements; implementors are the
only ones who can be expected to follow them.)

There's an additional problem: even if Web app authors wanted to avoid
depending on unspecified behaviour, in practice they cannot. There are no
tools to detect if a Web app depends on something unspecified. All they can
do is test in different browsers. If all the tested browsers happen to
behave in the same way, but other behaviours are allowed by the spec, the
author won't know this. Exhaustive manual analysis by an expert might find
some of the problems, but that's uneconomic.

Rob
-- 
"He was pierced for our transgressions, he was crushed for our iniquities;
the punishment that brought us peace was upon him, and by his wounds we are
healed. We all, like sheep, have gone astray, each of us has turned to his
own way; and the LORD has laid on him the iniquity of us all." [Isaiah
53:5-6]