[rfc-i] New version: draft-hoffman-utf8-rfcs-04.txt

Julian Reschke julian.reschke at gmx.de
Tue Nov 4 08:20:19 PST 2008


Joe Touch wrote:
> Julian Reschke wrote:
> ...
>> How is this relevant, though?
> ...
>> The problem is that after saving to a local file, the mime type
>> information is lost (including the encoding information), so what
>> happens after that solely depends on the operating system's treatment of
>> text files (which, at least for WinXP, has nothing to do what RFC2046
>> says about text/plain).
> 
> Agreed. This regards the use of the .txt suffix with a BOM to overload
> .txt to also mean UTF-8.

No, disagreed.

As long as operating systems do not store encoding information with the 
file, applications will always have to decide what encoding to use when 
reading the file. Some decide based on the user's locale, some guess 
from the file contents, some do not guess, but still peak for the BOM.

So, no, there is no "ASCII" default here.

> The BOM causes other problems (it changes the checksum of a file, etc. -
> see RFC3629).

How is that a problem when the original file (as obtained from the RFC 
Editor or the IETF) already has it?

BR, Julian


More information about the rfc-interest mailing list