Re: revised w/nocase text, considering titlecase and cased
John Cowan 10 May 2014 22:56 UTC
Alex Shinn scripsit:
> What I will do is specifically note that
>
> (w/nocase upper)
> (w/nocase lower)
> cased
>
> are all the same thing (where cased is characters with
> the cased (L&) property),
But they aren't the same thing; I already showed that. "t" with
diaeresis matches lower and so (w/nocase lower), but it does not match
(w/nocase upper). Nor do the mathematical capital letters, which match
upper and so (w/nocase upper), match (w/nocase lower). By my count there
are 1259 L& characters like this, so your three expressions match three
different sets.
I think the tangent on what Unicode should or should not have done has
distracted us from the original point. My recommendation is either
to add "title" explicitly (only 31 characters) or else just forget the
whole thing.
--
John Cowan http://www.ccil.org/~cowan xxxxxx@ccil.org
Female celebrity stalker, on a hot morning in Cairo:
"Imagine, Colonel Lawrence, ninety-two already!"
El Auruns's reply: "Many happy returns of the day!"