Rob Crowther wrote:

OK, then define "parsed and rendered correctly".

Exactly as you meant it in your earlier message :

whether or not a browser will correctly parse, interpret or display
it.


Or, put another way: where is the parsing process for a text file
conforming to HTML4's DTD defined so that we can judge the
correctness of a given browser's parsing behaviour?

The specification for the parsing process for HTML 4.01 is directly
derivable for the specification for the parsing process for SGML,
taking into account any notes in the DTD where the exact behaviour
could not be specified in SGML or differed therefrom.

Since the spec is based on what browser actually do, the only way it
will change 'at any time' is if all the major browsers suddenly
changed their behaviour.  If the spec didn't change at that time,
because it was a Recommendation or whatever, would you write your
documents to conform to the spec or to conform to what browsers
actually did?

To the specification, of course [1] : if I don't write to the
specification, I have no right to complain to the browser vendors when
their implementations fail to conform.  To base a specification on what
a particular subset of browsers do at some arbitrary point in time
is to completely fail to understand the reason for a specification
in the first place.

Philip Taylor
--------
[1] When current browsers (Netscape 4, I think it was) required four
non-standard attributes in the <BODY> tag in order to get things
positioned satisfactorily, all pages for which I was responsible
(a not-inconsiderable number) specified an augmented DTD in the
DOCTYPE directive in order that the pages both validate and render
correctly.  I am unaware of any facility for augmenting the HTML 5
non-DTD in a similar way to get around future failures to conform.
______________________________________________________________________
css-discuss [css-d@lists.css-discuss.org]
http://www.css-discuss.org/mailman/listinfo/css-d
List wiki/FAQ -- http://css-discuss.incutio.com/
List policies -- http://css-discuss.org/policies.html
Supported by evolt.org -- http://www.evolt.org/help_support_evolt/

Reply via email to