Lists Home |
Date Index |
Are the W3C implementations *reference* implementations
or *sample* implementations? IOW, how is the implementation
tied to the specification in terms of features and proof
of conforming and compliant implementation? Are any public source?
Are they tied normatively to the spec or informatively?
Anyone can code anything. That doesn't mean it works as a
goodness proof or can be used to test, validate, or verify
conformance and compliance. ISO is quite sticky about that.
Dick Puk gave us excellent advice on the topic and I am no
expert here. One will want to consult the relevant policy
<hearsay>We did a few rounds on the topic in the X3D process given
the stringent needs for client behavioral and rendering
fidelity (particularly, behavioral fidelity). AFAIK, reference
implementations don't have be performant but they must
strictly conform and comply. Sample implementations just
show that someone can implement the spec, but the meaningfulness
of it is subject to interpretation. If they cite an implementation
as THE reference implementation, that is the one on which
to base the decision, but again, read between the lines.</hearsay>
From: Simon St.Laurent [mailto:firstname.lastname@example.org]
At 12:17 PM 6/20/2003 +0100, Michael Kay wrote:
> > Maybe W3C should require the development of a reference
> > implementation for every spec. This is the way Sun's Java
> > Community Process works.
>W3C arguably goes beyond this: it requires multiple implementations of a
>specification before the spec goes from Candidate Rec to Rec status.
I would have infinitely more faith in that process - and the process by
which Candidate Rec is occasionally skipped - if there was a public report
acknowledging which implementations were used as the basis for the
decision. (And no, I wouldn't count member-private implementations as
enough to move a spec forward.)
Right now, there's no way to evaluate whether the W3C goes beyond the JCP
or IETF on this. My under-informed understanding is that it does not.