[erlang-questions] Scharfes S (WAS: Erlang 3000?)

[Richard O'Keefe]

On 20 Nov 2008, at 6:21 am, David Mercer wrote:

> There are two use cases of case-changing that I can think of.
> The other, however, is normalization, probably for comparison
> purposes––i.e., case-insensitive comparisons.  In that case, being  
> correct
> typographically isn't as important as consistency in how you  
> represent a
> character.  Ignoring the Unicode normalization approaches (regarding  
> accents
> and combining characters, etc.), a simplistic up-casing algorithm  
> may be all
> that is needed.

(1) The Unicode standard strongly recommends (with rationale)
     that *DOWN*casing should be used for such normalisation.

(2) If "straße" and "STRASSE" don't normalise to the same thing,
     for many millions of users, the normalisation is wrong.  The
     point at issue is that this is not a one-to-one mapping on
     >characters<.  (That is, there is no function F such that
     map(F, String) will give you an appropriately normalised result.)