Re: IRIs and bidirectional formatting characters

On Tue, Aug 25, 2015 at 6:31 PM, Phillips, Addison <addison@lab126.com> wrote:
> https://www.w3.org/International/wiki/IRIStatus

Looking through this, the URL Standard addresses the IDNA issues you
mention. It does not address how mailto is to be processed however,
but this could be build on top of what the URL Standard has today.

The bidirectional section has a good point in that even if we give
some advice towards displaying URLs (including converting their
internal percent-encoded bytes to Unicode) doing anything more than
left-to-right string is tricky and it might indeed be confusing if
browser UI displayed these differently from plain text. (This further
argues for only showing domains in the browser UI, I think, to avoid
making the path look like part of the domain.)

I'm not sure URL detection within a string of text is a worthwhile
standardization topic, but I suppose we could mention it and some of
its pitfalls. It seems though that determining whether n:m is a URL or
a relationship in a diagram is rather hard.

It's not clear why "Query encoding" is a subsection of "Link
detection" or why it's considered to be an open issue. The URL
Standard addresses it.

As for "Confusable characters" I think this is largely up to
implementations. I did add a section detailing some of the issues
here:

  https://url.spec.whatwg.org/#security-considerations


-- 
https://annevankesteren.nl/

Received on Tuesday, 25 August 2015 17:57:59 UTC