Status of recirculation ballot
(1) I think Vincent's "technical" comments should be considered editorial,
especially his r01.4 about non-negative -> natural, as the intent is
clear. It would indeed be technical if it was not contradicted by the
later Table 9.5.
(2) I actually disagree with his comment about 9.7.5, All-flavor decorated
interval literals. The decoration "com" is only required when more
than one flavor is supported. A single-flavor implementation can still
use all-flavor literals, but it would use its flavor-specific applicable
decoration if it does not support "com".
Table 9.5 does show that "com" has to be supported in the syntax, but
its value would then be the single flavor's equivalent. So I agree
that there is a discrepancy here.
So how do we fix this?
It *would* be technical change to require all flavors to support "com",
even in single-flavor implementations.
However, the intent of the all-flavor literal syntax is (I presume) to
have a flavor-neutral way to specify what in fact are common intervals, so
I suppose all flavors should be required to support "com" as a *name* for
whatever flavor-specific decoration corresponds to this. So Vincen'ts
replacement for 9.7.5 needs to be clarified.
Vincent had proposed:
An all-flavor decorated interval literal is a string s comprising
an all-flavor bare interval literal sx and the decoration literal
com, separated by an underscore "_". Its value is x_com.
I propose to change the last sentence therein to:
Its value is x_com, or, in
a single-flavor implementation that does not support "com", that
flavor's equivalent decoration.
(Then Table 9.5 can remain unchanged, and given that the table is
arguably the official specification, our suggestion is clearly just
an editorial clarification.)
Michel.
---Sent: 2015-04-14 14:24:45 UTC