Re: RFC 2822 email addresses in tag URIs

Frank Ellermann wrote to the URI Interest Group’s mailing list
(<mailto:uri@w3.org>) on 13 October 2005 in “Re: RFC 2822 email
addresses in tag URIs” (<mid:434E6B82.3EBA@xyzzy.claranet.de>,
<http://www.w3.org/mid/434E6B82.3EBA@xyzzy.claranet.de>):

> if you say "NO" to percent-encoding where necessary - 
> like always for UTF-8 at least in the LHS - then I'd think 
> that this could be a mistake.

When the time comes for the next specification of the “tag” scheme,
specify tags as IRIs that don’t need percent-encoding.

> And for some characters you
> need it anyway - I'm not sure, what about say "#" i the LHS ?

I came to the position that it is better to forbid certain e-mail
addresses in “tag” URIs than to allow percent-encoding within the
<emailAddress> portion.

> Sure, you don't need to support UTF-8 immediately.  But when
> it's available in some years it could be nice, e.g. the Czech
> example in the Jabber URI draft:
> 
>         <xmpp:ji%C5%99i@%C4%8Dechy.example/v%20Praze>
> 
> The part ji%C5%99i@%C4%8Dechy.example could be a future IMA.

The problem is that percent-encoding is not human-tractable in the case
of most humans. The solution, as stated before, is to make tags into IRIs.

-- 
Etan Wexler.

Received on Saturday, 15 October 2005 10:44:29 UTC