Re: percent-encoding in tel URIs

Clive D.W. Feather wrote:
  
> In other contexts, I've assumed that # was meant to be encoded.

If that does the trick it could be simple enough for an erratum.
Maybe there are other problems in the syntax, I didn't check it.

> The tel URI scheme is using <opaque_part> rather than any form
> with an <authority>.

Yes, my simplification was the general idea, no authority as in
"mailto", or optional as in "xmpp" is fine.  The rest still has
to match the general syntax.

> "?" does *NOT* start a query, it is an uninterpreted character.

In the general syntax it starts a query, for URI schemes without
query "?" cannot be used as is before the fragment (if there is
a fragment).  

I think, maybe it is wrong, a general parser not knowing scheme
specific tricks only needs to find the fragment, not the query -
and if that's the case I don't see why "?" cannot be used... :-(

> I'd be willing to update 3966; I take it this is the right WG?

In the RFC 3966 errata the author mentioned an "IPTEL WG", that
still exists.  Without <query> or <authority> most URI problems
for "tel" should be limited to <pchar> or not.  RFC 3966 is at
PS, 3966bis could go for DS, that sounds like fun.

 Frank

Received on Monday, 30 June 2008 13:03:58 UTC