Re: [cgmo-webcgm] for Wednesday telecon -- implementations

[Lofton Henderson <lofton@rockynet.com>]

At 08:37 AM 3/7/2005 -0500, Benoit Bezaire wrote:
>Sunday, March 6, 2005, 5:39:48 PM, Lofton wrote:
>[...]
>LH> Here is the problem.  Instead of having something like 4-5 implementations
>LH> that we expected, and that members committed to, we only have two that are
>LH> showing any significant progress.  We could face a situation in the near
>LH> future where insufficient implementation halts WebCGM advancement.
>Unless Ulrich and I are lying about our implementation status, which I
>doubt that is the case :-), we're not doing too bad.

Yes, and the TC owes you gratitude for that.  But the problem with the 
situation is that it forces your two implementations to be complete (and 
correct of course).  Which is both risky (to us) and unfair (to you).


>LH> This would be extremely unfortunate, given all of the hard work
>LH> that so many have put into this.
>Yes it would. What I suspect will happen is that the specification
>will be scale down a bit. For example, there are things for us that
>are a higher priority than others, and I suspect Ulrich to have his
>own list of priorities. The features that are high priority for both
>of us are not at risk, but the other ones may be at risk.

Exactly.  Which is why the other vendors have to get busy and start 
delivering on their implementation commitments.

>That's why
>we need the test suite ASAP, to be able to compare.

Right.  And also (similar to SVG's practice), followed quickly by a matrix 
of all test cases (rows) versus all implementors (columns).  We can then 
quickly spot and decide how to deal with rows without two green cells (drop 
the feature, assign "volunteers" to implement, etc).

-Lofton.