On 17/2/09 4:13 PM, "Erik Wilde" <[email protected]> wrote:
> Aristotle Pagaltzis wrote: >> * Erik Wilde <[email protected]> [2009-02-16 19:45]: >>> <feed xmlns="http://www.w3.org/2005/Atom"; xml:via="..."> >>> <link xml:id="feed1" rel="via" href="http://example.com/feed-A"; /> >>> <link xml:id="feed2" rel="via" href="http://example.com/feed-B"; /> >>> ... >>> <entry> >>> <source via:via="feed1"> >>> <link rel="self" href="http://example.com/feed-A"; /> >> Under what circumstances would the entry self link ever differ >> from the feed via link? What¹s the use of those IDs and via:via >> attributes? > > i don't think that these two URIs would ever differ. based on my > scenario (providing aggregated feeds as a way to specify queries across > a number of feeds), there are are reasons for the IDs: > > - even if the aggregated does not contain any entry of a feed it is > aggregating, the <link /> still specifies that this feed is contributing > to the entries available through the aggregated feed. this would be communicated via the @rel attribute, surely? > - in the query on the aggregated link, the IDs are used in the query > itself to specify the entries a client is requesting, such as "i am > interested in all entries from 'feed1' where the title contains the > string 'atom'." Hmm ... interesting. So you're not necessarily aggregating *all* entries from http://example.com/feed-A but instead possibly just a subset. So another aggregator who wants to carry this further shouldn't munge your id values? This raises a further wrinkle in that your ids, even in your own application space, would need to be globally unique. Eg: if your application is providing a "top movie reviews" feed-set and also a "geeky web talk" feed-set, you generally wouldn't find the same id values used in both (unless it's a top movie review found in some geeky web talk forum). I would suggest something else to carry your feed-criteria data .. <link rel="via" href="http://example.com/feed-A"; via:criteria="http://via.example.com/feed-A/criteria/1"; type="application/atom+xml" /> That way your criteria idents won't get stomped on by any further aggregation. e.
