On 2008-02-04 23:06, you wrote:
> Aristotle Pagaltzis wrote:
> > * Daniel Aleksandersen <[EMAIL PROTECTED]> [2008-02-04 
16:10]:
> >> Why does so many in this list fail to eralise that I am talking
> >> about the <feed> element’s <id> element?
> >
> > Ugh, sorry.
> >
> > Well, the point about atom:id being globally unique continues to
> > apply in either case.
> >
> > However, I’m not sure what the feed ID really identifies, nor can
>
> The Feed level atom:id is primarily useful as a logical group
> identifier.  For instance, suppose you have a collection of Atom feed
> documents representing a paged set.  Together, all of the entries
> comprise a single logical group commonly identified by the atom:feed's
> atom:id element.  If there was no atom:id element, there'd be no way of
> grouping entries from multiple feed documents into a single logical
> set.

Then why is this not written in any specifications?

> > I recall any compelling argument about its meaning, so I have no
> > idea when feeds should share an ID. (The fact that everyone
> > thought you were talking about entry IDs is a strong indicator
> > that no one has much use for feed IDs.)
> >
> > In the context of Atompub it is somewhat clear: when two feeds
> > are views onto the same collection, they should probably have
> > the same ID. However, even then it is not necessarily clear
> > when a feed other than the collection feed is a view onto that
> > collection rather than a distinct entity that happens to draw
> > entries from the collection.
> >
> > Regards,
-- 
Daniel Aleksandersen <[EMAIL PROTECTED]>

Reply via email to