Re: the discussion so far

[Thomas Bushnell BSG <tb@xxxxxxxxxx>]

"John.Cowan" <jcowan@xxxxxxxxxxxxxxxxx> writes:

> It's important to note the rather subtle definition of "canonical
> equivalence" in Unicode.  It's not the case that if two strings are
> canonically equivalent, a Unicode-compliant process MUST treat them
> identically.  Rather, a Unicode- compliant process MAUST NOT
> assume that another Unicode-compliant process will treat them
> differently.

I believe that a sufficiently fancy Scheme implementation should be
allowed to treat canonically equivalent sequences identically.  We
should not standardize in Scheme a differential treatment here.