Lists Home |
Date Index |
The case for X3D/VRML was persuasive. Applications that render rich
graphics and rich behavior in real time have to be solidly interoperable.
It's not good enough in computer-aided design or games, or even just
spinning models to be less than that. So the conformance tests are
part of the pathway to being able to display testmarks and trademarks.
This part of the ecosystem requires conformance testing of products.
That is the difference between something like X3D and this:
Probably a good product, but a product that only interoperates
with itself is a pretty risky investment these days. XSLT
is not always the right answer, and it is not an answer at
all if the question has the word 'interoperate' in it. We
learned that the hard way in VRML when a half dozen browsers
could render close but not quite, and behaviorally, were sometimes
miles apart. Tony Parisi called it "rendering and behavioral fidelity"
and those are the horns these applications get right or users are
From: 'Liam Quin' [mailto:firstname.lastname@example.org]
On Thu, Oct 14, 2004 at 03:46:31PM -0500, Bullard, Claude L (Len) wrote:
> The right answer might be to create the conformance tests and
> enable someone else to do the testing. That's what we're doing
> for X3D. The caveat might be that some applications need thorough
> conformance testing, some only need some testing, some need none.
That's what we do too, except that we don't brand products based on
their success at handling the tests. The conformance tests are really
to test the specification -- can it be implemented interoperably? --
rather than to test products.
Liam Quin, W3C XML Activity Lead, http://www.w3.org/People/Quin/