- From: <ned.freed@innosoft.com>
- Date: Tue, 04 Apr 2000 20:15:10 -0800 (PST)
- To: Masataka Ohta <mohta@necom830.hpcl.titech.ac.jp>
- Cc: ned.freed@innosoft.com, "Martin J. Duerst" <duerst@w3.org>, Harald Alvestrand <Harald@Alvestrand.no>, ietf-charsets@innosoft.com
> Some unregistered charset names may be not usable as a MIME charset > parameter value. What you call some random charset outside of a protocol has nothing to do with the registration process. > > Martin is completely correct here. See RFC 2278 section 3.1. It says that all > > charset registrations must note whether or not they are suitable for use in > > MIME text, based on what the MIME specification (RFC 2045) says about text. > Wrong. RFC 2278 requires: My dear sir, I wrote this part of the specification, so I think I know what it means. You can argue that the words aren't clear if you like (although many people have read this specification and you are the only one that has arrived at this bizzarre interpretation) but you cannot argue what the intent was. > All registered > charsets MUST note whether or not they are suitable for use in MIME. You are quoting out of context. The full paragraph reads: Registered charsets MUST conform to the definition of a "charset" given above. In addition, charsets intended for use in MIME content types under the "text" top-level type must conform to the restrictions on that type described in RFC 2045. All registered charsets MUST note whether or not they are suitable for use in MIME. In context it is quite clear that this is talking specifically about MIME text, not MIME in general. The suitability requirement is with MIME text. indeed, arguing that the suitability require applies to MIME in general makes no sense at all, since there are in general no suitability requirements that cover all of MIME. I'll take this as a sign that the text in this paragraph need to clarified in the next version. But you should take it as a sign that your reading of the specification is wrong and needs to be corrected. > From the interoperability point of view, three charset names > for three different encoding methods of JIS X 0213 is much > worse than 3 aliases of a single encoding method of ISO-8859-14. > If you think, in IETF, written rules are overridden by > misinterpretations of the editors of the rules, even though > the rules are legislated partly by public review, that's fine. Nobody else has indicated that they believe the written text says what you think it says. In fact everyone else I've communicated with has indicated that they don't agree with you at all. And I'm not the editor, I'm the author. And this was discussed at great length and in enormous detail, so the meaning others attached to it is equally clear. Ned
Received on Tuesday, 4 April 2000 23:28:24 UTC