On 10/11/04 3:52 PM, "Bob Wyman" <[EMAIL PROTECTED]> wrote:

> Eric Scheid wrote:
>> <origin href="..." type="application/rdf+xml" title="My RDF Feed"/>
>
> That's more expressive than RSS V2.0's <source/> since it provides
> type as well as URL and title, however, it still doesn't provide feed-level
> metadata such as "author", "copyright", etc. that are potentially vital to
> the proper interpretation of an entry.

ah, but now you do have the @href, so it won't be hard to go fetch it

except ... besides the roundtrip pain you might encounter the current feed
with changed meta-data (eg. a guest author), different from when the entry
was actually published. Life would be so much simpler with a SSFF with
discrete <entry> documents ;-)

> Some feed-level metadata is
> inheritable by entries, other feed-level metadata is important for
> understanding the context in which the entry was published or within which
> it was intended to be interpreted.

just how far do we push the need for context though?

>> If an entry is originally published in Feed-A, picked up by an
>> aggregator and republished in Feed-B, which is then picked up by
>> another aggregator and republished in Feed-C ... what should the value
>> of atom:origin in Feed-C be? Should it be the atom:id of Feed-B,
>> or Feed-A?
>
> I believe that an entry can only have one "origin". In the case
> above, clearly the "origin" is Feed-A. It might, for some applications, be
> interesting to note from where the entry was most recently copied, however,
> that would be something other than origin.

so an (unwritten?) rule of atom:origin is that when aggregating entries from
source feeds, if the entry already has an atom:origin don't replace it with
a new one?

e.

Reply via email to