[
Lists Home |
Date Index |
Thread Index
]
Elliote Rusty Harold wrote:
> How would you feel about dropping all language elements and
> replacing them with xml:lang attributes?
It is clear that xml:lang is vastly superior to the
<language> tag of RSS V2.0. xml:lang permits us to identify
language use with the greatest possible granularity. But,
what I was trying to do was figure out if there is any way to
address this problem within the constraints of the RSS V2.0
specification. Doing otherwise would be creating a new
specification and that is supposedly the task of Atom...
>an unexpected attribute is a lot less likely to trip
>people up than an unexpected element.
I'm not sure if that is actually the case. There seem to
be quite a few home grown RSS reading tools that don't use
XML parsers but rather grep for element tags directly. I'm
afraid that many of these things would not do too well
handling a tag with an unexpected attribute. In any case, the
only way to find out is to give it a try and see what
breaks... Not pleasant. Life would be easier if people would
simply learn to use parsers instead of mucking about with the
bytes directly...
In other mail, it has been pointed out that the ISO
language codes include "mul" (multiple) and "und"
(undetermined). Thus, what I was suggesting in my earlier
note (new tags i-unknown and i-mixed) is unnecessary.
However, RFC3066 specifically recommends against either mul
or und except in cases where there is no other choice. While
it might, on occaision, be "necessary" to use "und" even with
xml:lang, it seems to the use of "mul" would be very rare if
only because xml:lang allows you to change the language code
when needed. But, given the limitations of RSS V2.0, it might
be appropriate to use mul in that context. For instance, one
solution might be to set the RSS <language> tag to "mul" to
warn processors that the feed contains a mix of languages and
then use xml:lang tags to flag the specific items with their
appropriate languages where known. What do you think?
Thanks for the input.
bob wyman
|
