On Mon, Aug 28, 2023 at 6:29 AM Robert Stepanek <rsto=40fastmailteam.com@xxxxxxxxxxxxxx> wrote:
Hi,I'm still not convinced that there is something Bidi-specific to add to this specification, other than what Unicode standard and the current specification already define.
I am also not convinced, but perhaps some concrete examples might help. This technique worked well with addresses (which now look fine to me, although they might need refinements in the future). There, I can see some cases where the correct behavior is only implied, but nothing wrong.
I'm pretty familiar with RTL/LTR and whitespace, as Tweets are more difficult to write in many languages. They're much nicer to write in Western languages, since some entities (at-handles, hashtags, scheme-less URLs) appear in free text. The JSContact format is much more structured, so I don't really understand the concern here. Is there a specific i18n concern that is wrong? Not rhetorical, really asking.
We could take this example:
Does the spec cover it well? I think it's ok, although not very detailed. When I look at the definition:
"uri: String (mandatory). The resource value. This MUST be a URI as defined in Section 3 of [RFC3986] and updates."
It seems that "updates" could clearly allow this URL. How far does it need to go? I think the authors should decided the amount of ambiguity they're willing to tolerate.
thanks,
Rob
* 1988 Nobel Prize for Literature:
-- last-call mailing list last-call@xxxxxxxx https://www.ietf.org/mailman/listinfo/last-call
