Re: Interoperability/Compatibility

Shane McCarron wrote:
> Think of a standard as a covenant.  It says "If you, the portable 
> content developer, do things right, I, the content processor 
> implementor, promise that it will work."

These are the normal rules for standards.  However, one of the big 
issues here is that WHATWG's definition of these rules is different, and 
they argue, probably with some validity, that it is not possible to use 
these rules when web authors are involved.

The most important thing they do is to delete "If you, the portable
content developer, do things right" from the above, so that, whilst they 
will claim that they impose restrictions on authors, there are no 
consequences to the authors in violating those restrictions.

Possibly more controversialy, I believe that they define "work" as 
"produce what the author expected", rather than "process the input in a 
way consistent with its meanings as defined by the standards".  An 
example might be P. A non-empty P defines a paragraph, but IE produces a 
block of text followed by some vertical white space.  Authors expect 
every browser to behave like IE, so, if another browser indicates 
paragraphs by indenting the first line, I believe WHATWG would say the 
standard was not interoperable, because it allows one browser to produce 
the expected result and one to produce a different result.

(This does lead to a contradiction when they say that the reason to have 
more than one browser is that they can behave differently, as it is not 
clear why that difference in behviour won't also violate user expectations.)

Received on Tuesday, 1 May 2007 20:34:22 UTC