[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [idn] Extending gethostbyname()
- To: "D. J. Bernstein" <djb@cr.yp.to>
- Subject: Re: [idn] Extending gethostbyname()
- From: Keith Moore <moore@cs.utk.edu>
- Date: Tue, 29 May 2001 02:29:52 -0400
- cc: idn@ops.ietf.org
- Delivery-date: Mon, 28 May 2001 23:30:23 -0700
- Envelope-to: idn-data@psg.com
> Did the RFC 1123 authors worry about some badly written programs having
> trouble with names like 3com.com?
perhaps. but they did not specify that the APIs had to stay the same,
or for that matter, that they had to change. rather they assumed that
API writers would do something reasonable.
also, RFC 1123 didn't change the interpretation of the octets used
in a domain name. both before and after RFC 1123 those octets were
understood to be from the printable subset of ASCII. There was
no question as to whether a program printing the name obtained from
gethostbyaddr() before RFC 1123 would be able to print a name obtained
from that routine after a post-1123 upgrade, and the ambiguities
that crop up when using ASCII (eg 0 vs O, l vs 1, - vs _) were
reasonably well-understood.
These questions do crop up when considering whether to use native UTF-8
in IDNs.
Also, the 'net was *much* smaller and in many ways less diverse when
1123 was written. It was much easier to understand the likely impact of
such changes, and there were fewer constituencies that had to be satisfied.
Keith