On Tue, Jan 13, 2009 at 9:07 PM, James M Snell <[email protected]> wrote:
>
> As the author of 4685 I'm a bit biased but I'm not sure I see the need for
> the new rel attribute. in-reply-to would seem to cover the need well enough.
>

Just a thought in favor of in-reply-to:.  The "source" attribute could
be quite useful.  Per RFC 4685:

"""
   The "source" attribute MAY be used to specify the IRI [RFC3987] of an
   Atom Feed or Entry Document containing an atom:entry with an atom:id
   value equal to the value of the "ref" attribute.
"""

Seems to me that a client might be aware of a finite set of "sources":
YouTube, Digg, etc., and thus would be able to decide which
in-reply-to elements it would like to process.

--peter


> - James
>
> Bill de hOra wrote:
>>
>> Antone Roundy wrote:
>>>
>>> Martin Atkins wrote:
>>>>
>>>> I suppose the general case is that when the user adds additional
>>>> metadata their entry becomes a pointer to the original item rather than the
>>>> item itself, so the id (and everything else, for that matter) is different
>>>> in this case.
>>>
>>> That makes sense to me.
>>
>> Me too.
>>
>> The atom:source element isn't designed to capture that idea, so we're
>> looking at something along the lines of thr:in-reply-to or a new rel
>> attribute.
>>
>> Bill
>>
>>
>
>

Reply via email to