Stop me if I've misunderstood...

[Paul Hoffman]

At 3:42 PM +0900 7/11/09, Martin J. Dürst wrote:
>On 2009/07/11 8:28, Paul Hoffman wrote:
>
>> Also, I think many Asian user's would be surprised to hear that their IMEs don't map things like like half-width characters.
>
>Paul, what exactly do you mean here? Users would be surprised to hear ..
>because actually IMEs do map? Mapping from half-width characters?
>Or mapping to half-width characters?
>
>IMEs are extremely widely configurable. But that means different users
>will have them configured differently.

Yes, definitely. If we say an "IMEs SHOUD NOT map" to an implementer whose IME currently allows mapping, we either are saying:

- Change the way you do things and surprise the user

- Keep doing things the way you do now and surprise people who were relying on this interoperability

I recognize that not all IMEs map, and that those that do, do it however they feel like. This is why I am trying to drill down on what people who want to start using SHOULD and MUST actually mean. Pete and I could toss them back into the document, but the outcome will be surprise later when an implementer in good faith does something we did not expect.