From: Kenneth Whistler (kenw@sybase.com)
Date: Mon May 05 2008 - 17:20:33 CDT
> > I'm not seeing the useful clarification here. The conditions are
> > trivially true for any uncased character if you want to test the
> > edge condition and assume X = Y, since an uncased character casemaps
> > to itself.
>
> If you allow X = Y to be valid for application of the rule, uncased in
> version V then implies uncased in every version.
O.k., I see your point. I think this comes down to an instance
where the English expression of the principle was clear, but
where turning it into a logical statement caused mischief because
it does not rule out the X = Y case, even though most would consider
that nonsensical for a "case pair".
So the initial statement, "Two assigned characters form a case pair when..."
is highly unlikely to be interpreted by any but the most perverse
as applying to "two assigned characters that happen to actually
be the *same* code point and which could then be considered to
be a "case pair" only by lawyering it to death..."
But in the logical expression, it probably makes sense to add
the "where X != Y" restriction.
--Ken
> This is not what you want
> to convey in the guarantee, and moreover later on you just pointed out that
> it was not true. Having made this unintended deduction, the reader might
> then dismiss the statement as false - especially if the reader were
> jaundiced against Unicode, as some seem to be.
>
> Richard.
>
>
>
This archive was generated by hypermail 2.1.5 : Mon May 05 2008 - 17:23:18 CDT