When are and aren't two URLs the same? (ports)

[Drummond Reed]

I can't resist pointing out that this was one of the key issues we tackled
with XRIs. As identifiers they are designed exclusively for the digital
identity layer, so there are two types of XRIs (authority i-names and
i-numbers) that explicitly do not use port numbers and thus don't have that
normalization issue.

Also, we spent a bunch of time on normalization rules so comparison would be
as foolproof as possible.

However I believe the same problem can be solved at the URL layer by doing
the same thing: publishing strict normalization rules and requiring them.
And it should happen at the Yadis layer so that they are the same for all
URL/URI-based Yadis identifiers.

=Drummond 

-----Original Message-----
From: yadis-bounces at lists.danga.com [mailto:yadis-bounces at lists.danga.com]
On Behalf Of Johannes Ernst
Sent: Thursday, April 13, 2006 8:39 PM
To: Grant Monroe
Cc: Yadis list
Subject: Re: When are and aren't two URLs the same? (ports)

On Apr 13, 2006, at 16:57, Grant Monroe wrote:
> If IE isn't going to distinguish between the two URLs, we'd be
> fighting a loosing battle to do anything different.

I don't think IE's behavior has any bearing on the subject -- other  
than as an interesting analogy, which is the only reason why I  
mentioned it in my e-mail.

The particular interop problem that occurred did not involve a  
browser as a significant part. It occurred when the user (me) entered  
something like
     http://example.com/joe
into a Relying Party, who then went on to attempt to get the OpenID  
HTML document (could have been the Yadis URL, or the LID meta=lid  
query) from URL
     http://example.com:80/joe

So I think this remains a valid issue, and I'm sure we'll see lots of  
issues like that going forward as adoption of URL-based identity  
accelerates ...




Johannes Ernst
NetMesh Inc.