W3C home > Mailing lists > Public > public-html@w3.org > April 2007

Re: The argument for |bugmode|

From: Matthew Ratzloff <matt@builtfromsource.com>
Date: Fri, 20 Apr 2007 11:07:03 -0700 (PDT)
Message-ID: <49644.152.157.114.93.1177092423.squirrel@webmail.builtfromsource.com>
To: public-html@w3.org

On Fri, April 20, 2007 1:10 am, Lachlan Hunt wrote:
> This is why I am pushing so hard for the default to be always-standards
> mode.  It absolutely must be the default so that this really can be an
> interim solution for IE, not a permanent solution for everyone.

I've thought about this issue a great deal in the last few days, comparing
HTML to SGML, XML, and even spoken language.

I've decided that a version attribute and explicit opt-in are the wrong
way to go.  Defaulting to always-standards mode, with a "bugmode"-like
attribute, is the best method for moving forward.

HTML 5 introduces a new pseudo-DOCTYPE, "<!DOCTYPE html>", which is
already an implicit opt-in for HTML 5.  Even for vendors who want to keep
older rendering modes altogether separate, this DOCTYPE guarantees content
creators that their pages render in HTML 5.

Content creators must be made aware ahead of time of a switch to
explicitly peg rendering to a browser version.  I see this as an education
issue.  When and if IE 8, 9, etc. are introduced and potentially break
rendering in a small subset of pages, Microsoft (for example) can point to
the HTML 5 specification and say that there was always an option to opt
into a particular version's rendering mode.  The onus to educate the
public about this attribute is on those vendors wishing to make use of it,
and it must be mentioned openly and often leading up to the final release
of HTML 5.

Part of the anger at Microsoft, I think, stems from the fact that there
has never been an easy method for forcing page rendering to stay the same
between versions.  Conditional comments require more code on each page,
and for rendering issues to be resolved two different ways.  If IE 8 or 9
were to break pages, Microsoft (or any vendor) has an easy solution: add a
tiny attribute to your pages.  It's as easy as doing a find and replace.

Failure to make the default the latest standard will mean that every
single HTML 5 document in existence (and, yes, it will be every page) will
be forced to include information that very likely does not even apply to
them.  I believe it will also hinder the Web, as other browsers will
likely be forced to adopt this backwards-facing stance in order to
maintain compatibility with IE.  Furthermore, it encourages authors to not
depend on browser-specific hacks--which I feel hinder the web more than
anything.

-Matt
Received on Friday, 20 April 2007 18:07:18 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:43 UTC