W3C home > Mailing lists > Public > public-html@w3.org > September 2008

Re: Is "breaking the Web" with HTML 5 a non issue?

From: L. David Baron <dbaron@dbaron.org>
Date: Mon, 22 Sep 2008 11:36:09 -0700
To: Sam Kuper <sam.kuper@uclmail.net>
Cc: Justin James <j_james@mindspring.com>, Andrew Sidwell <w3c@andrewsidwell.co.uk>, public-html@w3.org
Message-ID: <20080922183609.GA5623@pickering.dbaron.org>

On Monday 2008-09-22 18:28 +0100, Sam Kuper wrote:
> ). But suppose for a moment that the browser vendors did play ball. What
> other problems might be caused by having the DOCTYPE do as it gives every
> impression of being supposed to do (i.e. specify the type of the document)?

One problem that would cause is rapidly increasing complexity of
browsers.  This would make it harder to write browsers (see
http://lists.w3.org/Archives/Public/public-html/2007Apr/0279.html )
and would drastically slow down the ability of browsers to implement
new features, because:
 * there would be fewer browser layout engines on the market (since
   it's harder to build them), and thus less competitive pressure
 * the browser layout engines being developed would be spending more
   of their time fixing security bugs, memory use bugs, and
   performance problems in their code (since they'd have more code),
   and thus be spending less time working on new features
 * implementing new features would require more work, since it would
   be associated with adding support for a new version of something,
   and making sure the feature was disabled in other versions.
These, together, would hurt the Web in its competition with closed
alternatives.

Another problem is that browser vendors are actually trying to make
their behavior more compatible.  This makes things easier for
authors, since they have to spend less time working around browser
bugs.  A major part of that effort involves making the
specifications clearer where they are now ambiguous.  HTML5 has made
many improvements to the precision and clarity of previous HTML
specifications.  If each version were handled according to its
specification, then browsers, in order to become more compatible
with each other (and, in many cases, with the content on the Web),
would want additional working groups to do further development of
HTML4, HTML 3.2, HTML 2, etc., so that the numerous ambiguities in
those specifications could be clarified, as they are being clarified
in HTML5.

Yet another problem would be that there are many documents on the
Web today that declare themselves as being one version of HTML, yet
use features from other versions.  Web browsers
that display documents on the Web well enough that users will use
such browsers must handle these documents, including the features
that are in the wrong version.  This would require, in order to get
interoperability, that the handling of those features be added to
the specifications (above) for HTML4, HTML 3.2, etc.

-David

-- 
L. David Baron                                 http://dbaron.org/
Mozilla Corporation                       http://www.mozilla.com/
Received on Monday, 22 September 2008 18:37:00 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:23 GMT