Re: [idn] IDN security and ACE leakage

[Keith Moore <moore@cs.utk.edu>]

> >the time required for MIME deployment has nothing to do with charset issues,
> >and everything to do with the large installed base of RFC 822 text-only
> >mailers,
> >and the design decision for MIME to provide for carriage of a tree-shaped
> >structure of arbitrary, labelled body parts.
> >
> >even if you think that supporting multiple charsets delayed MIME further,
> 
> Well, probably there was some synergy effect, so in part it may have
> made things faster. But there are still some mailers that do quite
> a good job on multiparts but a bad job on charset issues, so in part
> indeed things have been slower.

There are so many things required of a good MIME implementation that 
it's not surprising that an implementation does a better job with
some parts than others.  But again, this isn't the fault of MIME unless
you believe that MIME should have solved *only* the charset problem and 
not bothered with attachments and the structure thereof.
 
> >there was no universal charset available at the time that MIME was defined.
> >would you have delayed the work on MIME until 10646 was finalized
> >(several years later)?
> 
> I acknowledged that. But now the situation is different.

so is the problem.  However, the current IDN question and MIME do have one
significant thing in common - the need to avoid unnecessary disruption of
the installed base.

Keith