Re: Bidi issues

 > >       [email protected]
 > 
 > That looks shocking to me, but it's consistent with the IRI draft
 > section 4.4 example 5:

That was my first reaction, too, but I can understand why it might
make sense to a user of a computer in an RTL language.  LTR-in-RTL
embeddings happen every time a Hebrew or Arabic user quotes ASCII text
in an RTL passage.

So if NIAMOD@RESU _looks_ to the Hebrew/Arabic speaker as a single run
of RTL text, it's probably not that unnatural for them to get used to
reading this as [email protected]

A corollary of the above is that any Hebrew or Arabic package that
_allows_ ASCII text to be used will inevitably have some kind of bidi
algorithm -- even if only a simplistic one -- unless it is going to
require you to type the ASCII text backwards.  If we assume that --
over time -- the Unicode bidi algorithm is going to become the de
facto standard, then users of Hebrew or Arabic systems may indeed find
it natural (or at least unsurprising) that something they read as
[email protected] appears as [email protected]

On a side note, and to hark back to my hobby horse of a standard for
printable represenations of e-mail addresses:

It seems clear from the discussion that there is some scope for
disagreement as to whether the printed form

	[email protected]

should mean user FOO and domain BAR.com or user BAR at domain FOO.com.
It would clearly be madness if some MUAs chose one interpretation and
others chose the other.  Whether this should be addressed in the IMAA
document or a companion BCP is for others to decide, but I don't think
this is one we can leave to implementors' preferences...

	-roy