- From: Larry Masinter <masinter@parc.xerox.com>
- Date: Mon, 22 Mar 1999 19:44:55 PST
- To: "Martin J. Duerst" <duerst@w3.org>, "Jacob Palme" <jpalme@dsv.su.se>
- Cc: "IETF work on revision of RFC1036" <usenet-format@clari.net>, "TERENA Working Group on Internationalisation" <WG-I18N@TERENA.NL>, <imc-intl@imc.org>, <www-international@w3.org>
My (only casually considered) opinion is that adding more MIME headers for language translations won't help with the applications it is purported to support (content selection and content preparation), and that a more useful direction would be to register "language" and "translation" content features using the CONNEG framework, allowing the expression of content-features, recipient preferences, etc. in a uniform framework. Among other things, the conneg framework (RFC 2533) allows expression of 'q factor', and indicating some estimate of the likely "quality" of translation and acceptability seems to be a useful addition to the list ordering of multipart/alternative. http://www.imc.org/ietf-medfree/ points to the CONNEG working group items. Note also that some of the media feature response mechanisms in RFC 2530 would work for indicating language preferences and that language preferences for given email recipients could be expressed in the 'resource capabilities' expression (http://www.ietf.org/ietf/99mar/rescap-agenda-99mar.txt). Larry
Received on Monday, 22 March 1999 22:45:05 UTC