Reserved general punctuation

[Frank Ellermann]

Mark Davis wrote:

> I agree with Ken and John, after listening to this discussion.
> There is no really functional difference (as I said before), and
> I think the conceptually simpler option should take the field.

It's normally easy to convice me that "KISS" is better than more
convoluted solutions, but in that case I fear you go to far:

There are "unassigned" code points where we already know that if
they are ever assigned the result is guaranteed to be DISALLOWED.

When we already know this it would be nice to have it clear in
the tables once and forever, instead of other mechanisms used to
classify "not more UNASSIGNED" code points.

The tables draft is not only a snap shot of an enumeration like
say 4645bis, it comes with its own rules, and those rules are
not mere pointers to corresponding definitions in TUS 5.y, they
really do something that is not directly visible in TUS x.y.

 Frank

P.S. for Ken, U-label u+7cba u+7b80 was supposed to be a real
example for the xn--cocacola class of A-labels.  No plausible
definition of DISALLOWED would catch it, but making DISALLOWED
as big as possible could still help.