Re: Comment from CSS WG on HTML DOM draft

I'm not opposed to adding the "indeterminate" flag. Personally I'd prefer a
single tri-state value rather than one bit of data and one bit of "is it
meaningful", for clarity and coding convenience,, but if the current value
is a boolean I agree that changing it to a short integer or bitfield would
have backward-compatability problems.

(However, one minor quibble: I'd argue that IE for Mac and IE for Windows
should be treated as two ports of one implementation, not an example of
existing interoperability/standardization. This certainly doesn't mean we
shouldn't adopt their solution if it's a good one; it just means that as a
precedent it has less weight than if multiple vendors had already
implemented it. And yeah, I'd say the same thing if it was IBM's
implementations.)

______________________________________
Joe Kesselman  / IBM Research