[[ This message got lost in the traffic on optional characters, but
is as important of a change to the document. Comments are welcome. ]]
>Greetings. After the discussion in Pittsburgh on the nameprep draft,
>Marc and I have an open issue on when to prohibit characters.
>
>The current document has the order of processing looking like:
>1) Check for prohibited input (many)
>2) Fold case
>3) Canonicalize with normalization form KC
>
>A possible alternative has been proposed that adds a step:
>1) Check for prohibited input (a few, just for case)
>2) Fold case
>3) Canonicalize with normalization form KC
>4) Check for prohibited output
>
>I believe that the proposed alternative is a good one, even though
>it complicates the processing some by having two lists. The
>advantage is that it would disallow fewer characters on input, and
>therefore cause less surprise for users.
>
>
>--Paul Hoffman, Director
>--Internet Mail Consortium
