Q1 is mapping on lookup permanent or transitional?

[Paul Hoffman]

At 12:05 PM -0400 3/31/09, Vint Cerf wrote:
>Q1: Should the proposed mapping on lookup in a revised IDNA2008 
>protocol specification be a permanent feature of the protocol or 
>should it have a finite lifetime? Should it be required or optional?

It needs to be permanent, unfortunately. We need to split out "needed for backwards compatibility with IDNA2003" and "needed for sensible use of pure IDNA2008". Mapping is needed for the latter regardless of the former. That is, telling the world that "you can enter eurocafe.com and eurocafé.com, but not EuroCafé.com, in your browser" will have predictable and bad side-effects, such as some registries putting in characters that are forbidden. It sounds even sillier to say "you can enter eurocafé.com and EuroCafé.com until 2015, but at that time you must stop using the second one".

The question "Should it be required or optional?" is too vague to answer. Having a single type of mapping is required for interoperability of inputs (again, ignoring backwards compatibility with IDNA2003). If the current WG prefers the advantages of interoperability over the disadvantages of complexity as much as we did for IDNA2003, it is required. If the current WG cares less about interoperability than complexity, then it can be optional. Personally, I favor interoperability over avoiding complexity, and I think others here do as well; otherwise, we would not have adopted the complexity of the contextual rules.