Re: Back-references and USEAGE

From: Bruce Lilly (blilly@erols.com)
Date: Sat Jun 05 2004 - 15:36:50 CDT


Eivind Tagseth wrote:
> * Matthias Andree <ma@dt.e-technik.uni-dortmund.de> [2004-06-05 15:47:01 +0200]:

>>The other question is if a newsreader should localize headers,
>>particularly if the headers are editable in raw form. I believe it
>>shouldn't or it should at least offer an option (button, whatever) in a
>>PROMINENT location to show the headers as they are going to be sent so
>>the skilled user can check what his newsreader does, without requiring
>>support or test postings.
>
>
> No, I agree, a newsreader should not localize the content of an unstructured
> header. This illustrates the problems of enforcing semi-structure into
> unstructured content.

The field name is a protocol element ("name") not "text" [see RFCs 1958 and 2277]
and therefore isn't (or shouldn't be) subject to localization; only "text" has a
language; protocol elements do not.

Field body content of an unstructured field, however, *is* "text" and therefore does
have an associated language (RFC 2277). It is therefore not unreasonable to expect
that at some time, it may be possible to provide machine translation for the benefit of
users (e.g. babelfish -- and I would add that that technology has progressed rapidly
in the past 7+ years that this WG has been active, and there is no reason to preclude
the possibility w.r.t. this group's standard(s)).

Yes, it would be nice for software that provides translation (or decoding of encoded-
words, etc.) to also provide an option to display content as it will be transferred
"on-the-wire", but that's really a user-interface issue that ought to be outside of
our scope ("Internet standards normally do not specify the behavior of user
interfaces" -- RFC 3798).

Certainly we can envision difficulties with machine translation, but that should really
be outside of the scope of our discussion -- we should be dealing with format and
protocol issues and not user interface and presentation issues. Would the Chair
care to make a ruling on the scope of such discussions?




This archive was generated by hypermail 2.1.7.