Re: [idn] UTF-8 / RACE

[Patrik Fältström <paf@cisco.com>]

--On 01-05-27 22.24 -0700 Sherin Alsoudani <sherinalsoudani@hotmail.com>
wrote:

> Sadly, you and many people on this list are not understanding what I have 
> written.  My point has been to be practical: you say that the enduser can 
> control whether he is seeing the ACE encoded word or not.  Please show me 
> one client that can do this today.

There is no client which do UTF-8 encoding according to what this wg will
agree on, and there is no client which do ACE according to what this wg
will agree on.

The application the user is using have to be changed regardless of what
solution is chosen.

But, if you choose UTF-8 as the path, you have to also replace software in
the middle, which the enduser doesn't control, and which the enduser can
not change.

The enduser because of this have to wait until his network administrator do
the change.

In your environment, on your network, it seems that the network
administrator is doing changes fast, but that is certainly not the case in
many cases in the world (including at some sites in Arab countries I have
been working with).

> And after you find that one client, explain to me why I should change my 
> position regarding IE, given that IE is successfully putting out UTF-8 
> characters today, there are providers compatible with this today, and
> that  bringing the Internet to people sooner rather than later is
> important  *today*.

No, it is not doing it correctly.

See previous argument about nameprep.

Also, the Internet is not only the web, and no, a user can not communicate
only with a client with other peoples clients. Server software also have to
be replaced.

> I am puzzled by this amount of theory that is passed back on forth on the 
> list.  Things ARE moving, and UTF-8 is recognized as a better ultimate 
> solution.

I am not talking about the ultimate solution. I am talking about getting
started with IDN, and giving the users the power of doing so.

> Actually, you misunderstood: Arabic does not have any case - there is no 
> concept of upper/lower case.  But I certainly appreciate the need for
> what  you call nameprep; there are other issues that Arabic will need.

I know Arabic doesn't have any case. I know how it works. I promise you. I
wrote a directory service in 1995 with full UTF-8 support, especially for
Arabic.

But, the glyphs have different codepoints in Unicode. That was my point.


   paf