W3C home > Mailing lists > Public > www-tag@w3.org > February 2009

Re: Uniform access to metadata: XRD use case.

From: Eran Hammer-Lahav <eran@hueniverse.com>
Date: Mon, 23 Feb 2009 14:10:55 -0700
To: Jonathan Rees <jar@creativecommons.org>, "connolly@w3.org" <connolly@w3.org>
CC: "www-tag@w3.org" <www-tag@w3.org>
Message-ID: <C5C850DF.1310E%eran@hueniverse.com>
> From: Dan Connolly <connolly@w3.org>
> Date: February 19, 2009 12:56:21 PM EST
> 
> On Mon, 2009-02-09 at 16:32 -0500, Jonathan Rees wrote:
>> Pursuant to ACTION-200 (XRD use case), requested by Dan C:
>> 
>> http://www.w3.org/2001/tag/doc/uniform-
>> access-20090205.html#cross_site
>> 
>> This is rather quick and dirty, so let me know how you'd like to see
>> it improved.
> 
> It makes sense, though you correctly " anticipate that some will
> object
> that the information should have been put in the representation (i.e.
> found via GET)"

First, not all representations are capable of embedding such metadata (i.e.
video, audio, etc.). Second, there are as many people who find it
objectionable to mix data and metadata within the same representation.
 
> also... "This is the approach taken by openid 1.0, which failed to get
> the uptake desired" really? I read just last week about usability
> work that lead to some 80% success rate at google or something.

The 80% success rate (actually 92%) [1] has nothing to do with this. It is
the rate in which people who clicked on a Google button returned back to
Plaxo. This used what is known as Directed Identity in OpenID which
completely bypasses the entire flow Jonathan described...

The comment about OpenID 1.0 isn't accurate. OpenID 1.1 (the first real
interop version) used HTML <LINK> elements to convey its configuration
information. This is still supported in OpenID 2.0 and is widely used by
ultra-geeks who wants to have a "vanity-URI" (their own domain name).

Yes, they can as easily use XRDS for this (linked to from the blog or web
page), but they choose not to because they consider HTML elements to be
easier. I strongly disagree but that is not the right place for that.

However, this use of HTML <LINK> elements is an interop nightmare and causes
many sites to fail. For example, most OpenID sites use regular expressions
to scan the HTML page for those LINKs, forcing them to look for a specific
@rel value and break if there are multiple (space delimited) values in the
attribute.

EHL

[1] 
http://www.readwriteweb.com/archives/comcast_property_sees_92_success_rate_o
penid.php

 
Received on Monday, 23 February 2009 21:11:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:48:12 GMT