(cross posted to the atom-syntax list)
James Holderness wrote:
> [snip]
> So if you want to go with the bare minimum, then you should really be
> limiting yourself to a dir attribute on descriptions elements and nothing
> else. If you aren't going with the bare minimum then I'd argue for a
> consistent application of all form of direction encoding. I think it's
> clearer and less confusing.
>
Ok, I can understand the argument and although I can't imagine any real
world use cases, I can see how the lro and rlo values could be used.
The only harm in allowing them is a bit more complexity and an
additional requirement on the part of readers displaying the text (e.g.
that they treat it as an explicit override. I can go ahead and add
these to the proposed definition of the atom dir attribute.
However, I'm wondering how this could work in RSS given that we,
apparently, cannot have extension attributes in RSS. (I would, btw, be
very interested in hearing Dave's opinion on whether or not extension
attributes in RSS are allowed). Perhaps a "bidi:dir" extension element
on the channel and item elements could be used?
e.g.
<rss>
<channel>
...
<language>he</language>
<bidi:dir>rtl</bidi:dir>
...
</channel>
</rss>
It's imperfect, of course, but that's pretty much par for the course.
- James
