Backwards compatibility (was Re: [errorHandling-20] CLOSED...)

On Tue, 2003-12-02 at 10:45, Chris Lilley wrote:
> You also raised the issue of conformance to deprecated features, and
> suggested it might be a different issue. It seems that a deprecated
> feature is one that content authors and authoring tools should not
> produce, and that content consumers must understand.

Here's the problem I have with this.  Follow this timeline:

1998 - Person A writes something of profound geopolitical significance
(profound.html), in fully compliant HTML, per the modern specifications,
labels it "text/html", and sticks it on a webserver which is presumed
will run for the next 50 years.  The document goes largely
unappreciated...it's ahead of its time.
1999 - Person A gets hit by a bus.  She won't be revising her webpage
anytime soon.
2000 - W3C issues a new specification which uses the "text/html"
MIME-type, and deprecates many features found in profound.html
2003 - W3C issues yet another new specification which removes many of
the deprecated features (hey, they've been deprecated for years...who's
going to miss them).
2005 - Person B has a browser which fully complies with modern
recommendations for handling the "text/html" document.  They encounter
"profound.html", but their browser throws obscure error messages and
refuses to render the HTML.  Person B assumes that the document is
corrupted, and doesn't read the document.

This strikes me as a *very broken* architecture.  If anyone has
something of profound importance, they shouldn't need a time machine in
order to make sure that their work can be read in the future.

The W3C should stand by its recommendations.  If a particular document
type has accumulated too much cruft over the years (like, I don't know,
say "text/html"), then the document type should be discarded, and a new
one used.

Now, I realize that my characterization of HTML is not entirely fair.  I
have read the xhtml-media-types note [1], and understand that there's an
explicit caveat about using "text/html", but that's about helping new
content interoperate with old browsers.  It seems to be silent about
recommending that browsers claiming the "text/html" media type should be
prepared to deal with valid HTML 4.01 as a minimum.

I think the criticism can also be leveled when talking about the delta
between XHTML 1.0 and later versions of XHTML which remove support for
some elements.

Backwards compatibility is important stuff, and shouldn't be glossed
over so glibly.

Rob

[1] http://www.w3.org/TR/xhtml-media-types/

-- 
Rob Lanphier, Helix Community Coordinator - RealNetworks
http://helixcommunity.org http://rtsp.org http://realnetworks.com

Received on Friday, 12 December 2003 01:46:16 UTC