final sigma and tonos

[John C Klensin]

--On Thursday, 31 January, 2008 16:42 -0800 Erik van der Poel
<erikv at google.com> wrote:

> On Jan 31, 2008 4:25 PM, Kenneth Whistler <kenw at sybase.com>
> wrote:
>> Greek is not a cursive script, and ZWNJ has never had anything
>> to do with selection of final or non-final forms for Greek
>> sigma. ZWNJ is *not* a final form variation selector.
> 
> Is there any invisible character in Unicode that *could* be
> used by an application for whatever purpose it wishes?

Erik,

I appreciate the path you are trying to go down as well as Ken's
careful responses.  I've had similar thoughts about some sort of
presentation modifier (and mentioned one of them on the list
some days ago).  But all of them collapse on the same problem:
such a modifier helps us if the server knows that a string
containing it and one that does not still match.  It actually
probably helps us a lot more in some other situations than it
does in this one.  But we can't cross the line into having the
servers understand any of these things without discarding the
fundamental "no DNS modifications" goal for IDNA.  

So we end up back at the same point: if we believe that two
codepoints represent the same character, we need to map them to
the same character, discarding the information that they were
different.  If we believe they represent different characters,
then we need to be sure that they are not mapped together, lest
the difference disappear.   And if we got a valid form wrong in
IDNA2003 and Stringprep, then either it is wrong for all time or
we need to make an extremely disruptive change, such as changing
the prefix and creating a situation in which IDNA2003
implementations cannot even find names registered under IDNA200X.

      john