Let me say it again: the atom-bidi draft does not replace the use of the formatting characters. There are cases where the formatting characters will still be needed. I never said using the unicode characters was not an option; just that they weren't recommended. The *only* thing the atom-bidi draft does is provide a way of specifying the default direction of directionally neutral text. It does not attempt to replace the functionality of the bidi formatting characters. The bidi draft does provide a markup alternative to using the bidi formatting codes when all that is needed is the ability to specify a default direction for all the text in the feed, entry, etc. If you need anything more complicated than that, then you fall back to the unicode characters, which is not recommended but is still an option.
- James David Powell wrote:
[snip] Do you have a better way to implement bidi embedding in attribute values or PCDATA elements? The atom-bidi draft does not support bidi embedding. If you are representing bidi text in a scope that supports internal markup you can either use Unicode or mixed content markup; and mixed content markup is the prefered choice. If you are representing bidi text in a scope that does not support internal markup - which is what we are discussing here - then Unicode is the only choice.
