Re: comments on the comments...
John,
> You have shown my assertion (*) above is shaky. OK, there's a design decision here for xx = textToInterval(s).
> If we allow "promotion" from bare to decorated, I favour the standard saying
> "If xx's bare value is common, and context requires it to be decorated,
> the decoration shall be com in all cases."
>
> That is simple to understand. Maybe we already almost say this already, I'll need to check. Any objections to it?
The specification in subsection 9.4.5 "Portable decorated intervals" uses newDec(x).
Is section 9.8 "Operations on/with decorations" a proper place to formulate this property of newDec ?
-Dima