Re: [idn] Editorial comments on stringprep

[Erik Nordmark <Erik.Nordmark@sun.com>]

> I can't think of anything that would lead to that impression, in Unicode
> or anywhere else.  Uppercase, lowercase, and titlecase mapping are all
> handled on a completely equal footing in Unicode.

I didn't mean "undefined" as poorly specified.
Sorry for my lack of Unicode clue - perhaps I've misunderstood something.
But I recall hearing that given two arbitrary Unicode code points X and Y
one could do a comparison of
	MaptoUpperCase(X) and MaptoUpperCase(Y)
or
	MaptoLowerCase(X) and MaptoLowerCase(Y)

Both of these are well-defined, but the results of the comparison
will be different in some cases. I haven't verfied this but I think
U+0131 and U+0049 would compare as equal in the first case
but not in the second case.

But I now realize that the above isn't germane to the text on which I
commented. The above is just an argument that the spec must pick one or the
other techniques for case insensitive comparison.

Hmm - I don't recall if Patrik has some additional arguments for why
lowercase is better in his plenary presentation in Minneapolis.

  Erik