W3C home > Mailing lists > Public > www-tag@w3.org > October 2003

Re: Action item on syntax-based interoperability

From: Chris Lilley <chris@w3.org>
Date: Sun, 26 Oct 2003 22:28:33 +0100
Message-ID: <1448999250.20031026222833@w3.org>
To: Elliotte Rusty Harold <elharo@metalab.unc.edu>
Cc: "Dare Obasanjo" <dareo@microsoft.com>, "Tim Bray" <tbray@textuality.com>, www-tag@w3.org

On Thursday, October 23, 2003, 1:48:06 AM, Elliotte wrote:


ERH> At 3:59 PM -0700 10/22/03, Dare Obasanjo wrote:


>>So JavaScript and DHTML is not an interoperable part of the Web?

ERH> Of course it it's not, as anybody who uses anything other than 
ERH> Internet Explorer on Windows knows all too well.

In fact, people who have used *only* IE/win also find there are
interop problems.

ERH>  It isn't even close
ERH> to being an interoperable part of the Web. The sites that use syntax
ERH> work. The sites that use DOM and JavaScript don't.

Well, thats a bit of an exageration; as Dan Connolly said 'there
are additional risks' but to say tat all sites with DOM and
JavaScript/JScript/ECMAScript do not work is wildly overstating the
case.

The issue here is not that DOM is an API; the issue is that the spec
for the HTML DOM was developed some years after actual deployment, and
was designed to capture what people agreed (including te early
implementors) that things should move towards. It fails because the
move towards what the spec said was patchy and incomplete.

The XML DOM on the other hand is substantially more interoperable,
partly because there is agreement about what parsing means and partly
because the spec was produced ahead of the implementations, benefited
from early experience, and had a CR phase.

Its not whether something is a syntax or an API; its whether something
is well enough designed, quickly enough designed, and has enough
parties on board all going in the same direction implementation-wise.
And whether there is  CR period, a public test suite, and publicly
published results from named, version-numbered implementations to say
how they do on each test.

I therefore assert that interoperability is based on QA, not on
whether its a syntax or an api.

>>  Is
>>Flash also not part of the Web?

ERH> Ditto, though Flash is marginally more interoperable across platforms
ERH> than JavaScript and DHTML. That probably has something to do with
ERH> Flash not being designed by Microsoft.

Cheap shot. Microsoft make good stuff and bad stuff, like everyone
else.

The interoperability of Flash is due to

a) there being a single implementation from a single company (makes
interoperability trivial, I guess)
b) conformance in the specification is defined by 'what the latest
flash player from Macromedia does' in the case of a difference between
the spec and the implementation
c) success in getting the market to move forward to each new version
of the player. Incompatibilities between versions, which certainly
exist, is limited in effect because people auto-upgrade regularly. The
only old players around are those in other tools that picked up Flash
at one point in time (Java media, quicktime, etc have Flash 2 or 3
players)

The downside is of course that 'Mobile Flash' is based on Flash 3 or
thereabouts, different ActionScript, requires a different authoring
tool to desktop Flash, and has little interoperability with it.


-- 
 Chris                            mailto:chris@w3.org
Received on Sunday, 26 October 2003 16:28:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:22 GMT