[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [idn] I-D ACTION:draft-ietf-idn-uri-00.txt
- To: idn@ops.ietf.org
- Subject: Re: [idn] I-D ACTION:draft-ietf-idn-uri-00.txt
- From: "D. J. Bernstein" <djb@cr.yp.to>
- Date: 11 Jan 2001 01:07:41 -0000
- Delivery-date: Wed, 10 Jan 2001 18:01:26 -0800
- Envelope-to: idn-data@psg.com
- Mail-Followup-To: idn@ops.ietf.org
Adam M. Costello writes:
> UTF-8 gets less than 6 bits per octet on average for the non-ASCII
> octets, and the ACEs generally use base-32, which is 5 bits per
> character, so that's not much different.
ACEs also include several bytes to distinguish themselves from existing
ASCII names.
If you subscribe to my publicfile mailing list, you'll receive email
asking you to confirm the subscription. The Reply-To in the email will
look like this:
publicfile-sc.942884323.agghmppopdgakhmfhbaf-amc=cs.berkeley.edu@list.cr.yp.to
Now, what happens if your software rejects mail for usernames longer
than 64 bytes? publicfile-sc.942884323.agghmppopdgakhmfhbaf-amc=.xxx is
already 53 bytes, and the number is a timestamp that will soon increase
by another digit, so you have only 10 bytes for the rest of the name.
If you wanted an IDN, what would it be? Would it fit into 10 bytes with
ACE, not counting the .com?
> ASCII names are allowed to use mixed case. It would be unfair to deny
> the same flexibility to non-ASCII names.
I'd be happy to get rid of uppercase ASCII too. However, that would
require a transition period. In contrast, we aren't saddled with
existing uppercase IDNs.
---Dan