Re: [idn] A case for *ACE, other good/bad ideas, and a rant...

["Carl S. Gutekunst" <csg@eng.sun.com>]

> > Unless someone takes over the Internet again, there will continue to
> > be systems running cc: mail or worse.
> 
> Are you saying that cc:Mail isn't 8-bit clean?

cc:Mail could best be described as "8-bit erratic." An instance of the cc:Mail
client inherits the default charset for the platform on which it's running,
which might mean PC-437, PC-850, Shift-JIS, or whatever. This was an artifact
of MS-DOS that carried over to the MS-Windows client, and as a conseqeunce the
Windows client does character conversions between Windows 8-bit charsets and
the old PC charsets. These conversions are very dirty and irrecoverably munge
the data.

Shift-JIS is a great challange in its own right. Raw UTF-8, when passed to an
8859-1 application, is legal and replayable even though unintelligible. Not so
with Shift-JIS; a UTF-8 stream will contain many many code sequences that are
illegal in Shift-JIS, causing a well-behaved client to discard the data.

All of the native-language Japanese clients I know of also have this problem.
They use Shift-JIS natively, and will thoroughly destroy UTF-8 streams.

> Are you also saying that cc:Mail has no trouble with long domain names?

The length of Internet domain names when presented to cc:Mail is a complex
issue, depending on how the SMTP gateway is configured and whether the
recipient is listed in the cc:Mail directory. As a rule of thumb, however, the
limit should be the size of a cc:Mail postoffice name, which on older (but
still widely used systems) is 64 bytes.

<csg>