Re: <id> and <link> are broken! (was RE: base and atom:id)

[Tim Bray <tbray@xxxxxxxxxxxxxx>]

On Feb 26, 2004, at 7:28 AM, Dare Obasanjo wrote:
> The GUID identifies the post so if characteristics of
> the post change but not the post itself the aggregator
> isn't confused and treats it like a new post. Examples

Sorry, I still don't get it.

> 1.) I'm subscribed to an aggregated feed such as
> http://blogs.msdn.com and some individual feeds from
> that blog as well. With GUIDs, once I read an
> individuals post in either the aggregated feed then it
> can be marked as read in their individual feed as well
> or vice versa.

Why can't the aggregated feed identify the items by the same URIs that 
the individual feeds use?  I think that would be good practice.

> 2.) I'm subscribed to a blog and the links change
> either due to the blog moving or the user changing
> something about their weblogging tool. With GUIDs the
> aggregator isn't confused by new permalinks in the
> feed so it doesn't display all the already seen posts
> as unread.

Two distinct cases here; let's see if I understand:

- the blog itself has a new URI... so if I moved ongoing from tbray.org
  to textuality.com and I had GUIDs, then when someone repointed the
  aggregator from the old location to the new one, on the single 
occasion
  when the aggregator went there the first time, it wouldn't see the
  entries as new entries; assuming that it had a global database of all
  the GUIDs it had ever seen from anywhere.

- I change the 'ongoing' software so that the URIS change from
  /ongoing/200x/2004/02/22/xx to
  /ongoing/2004/02/22/xx
  then when the aggregator comes around for the first time after I make
  that change, once again it won't mark all the entries as new.

Right? -Tim

Attachment: smime.p7s
Description: S/MIME cryptographic signature