Re: UTF-7 is dead

From: Markus Kuhn (
Date: Tue May 25 1999 - 20:05:23 EDT

rch@WriteMe.Comwrote on 1999-05-25 23:25 UTC:
> AFAIK sendmail still can not handle 0x80-0x9F bytes
> in headers. Though 8bit bytes in headers is not allowed by RFC's,
> it.s widely accepted practice.

Agreed. And for headers, UTF-8 wrapped into base64 as you used it nicely
in your message seems to be perfectly appropriate to me. No reason to
introduce an extra hack such as UTF-7 that doesn't really fit into the
MIME distinction between "charset" and "encoding". Size really does not
matter much in headers, as we are only talking about single short lines
here anyway and not about bulk text. (Actually, size doesn't even seem
to matter in bulk text either: the difference between UTF-2 and UTF-4 is
eaten up by compression, or even just Moore's law every 18 months. The
bandwidth plague of the Internet today is multimedia content and not the
comparatively tiny amount of plain text that is transmitted between the
huge volumes of JPG data, RealAudio data, and blown-up zero-filled
Word97 doc headers.)

UTF-7 is dead and is today only implemented by a few large vendors who's
managers desperately have to find ways to keep their oversized
programming teams busy and therefore prefer monstrous feature fetish
over a careful simple-is-beautiful design. Nobody really needs UTF-7.


Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at,  WWW: <>

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:46 EDT