Re: [idn] IDNs in email message bodies

["Maynard Kang" <maynard@pobox.org.sg>]

> If you remove yourself from the concept of "text" but think of it as a
> "presentation" layer issues, then you will realise both of you are
> speaking of the same thing. :-)
> 

Perhaps I should have explained more clearly in my presentation (no pun intended) why I believe this problem is not simply a presentation layer issue.

Prima facie it looks like a simple problem: Applications would just hand everything off to the presentation layer to handle what's being sent on the wire, and the presentation layer would handle whatever that's received on the wire and present it in a form that is acceptable to the application.

It would be just a simple presentation layer issue if IDNs in existing application protocols were used only for resolution purposes (e.g. ftp, ping). 

However, some IDNs may appear in these protocol exchanges as placeholders or arbitrary data, and this means that the application protocol itself needs to know how to deal with an IDN. Henceforth it becomes an application layer problem too.

I heard the comment during my presentation that there were several layer violations, and rightly so. It was not my main intention to present any proposed solution to this problem, but rather, to point out that if we as a working group fail to set the direction on this issue correctly, there WILL be more layer violations in the future during implementation and, as the growing threat of "testbed" deployment looms, this will have many disasterous consequences if we do not deal with this now.

maynard