RFC4685 can be made to work for generic child relationships but it is rather awkward and the semantics don't fit well. I have often felt that a set of "up", "down" link relations would be helpful for a variety of use cases and would definitely welcome any effort to define such.

- James

Peter Keane wrote:
On Sun, Sep 28, 2008 at 5:49 AM, James Holderness <[EMAIL PROTECTED]> wrote:
Peter Keane wrote:
Wouldn't such an "up"  link relation beg the question of a link
relation to express a nested  feed (a al Google's "feedLink" extension
-- http://code.google.com/apis/gdata/elements.html#gdFeedLink?).  Such
a relation expressing a nested feed would be quite useful (a topic
which has been explored here a number of times before).
Linking to a feed can be achieved perfectly well using the existing link
element (see RFC 4685 for an example).


RFC 4685 includes the "replies" link relation, which is a more
specific example of what I was getting at.  While the Threading
Extension seems to be quite useful, it *is* quite specific, and I have
wondered if a more generic mechanism might be useful as well.
Honestly, as much as being about the need for a new "rel" value, it
gets back to the issue of being able to add metadata to a link element
-- RFC 4685 offers the thr:count attribute (analogous to the Google
feedLink 'itemCount' attribute) -- one of the very useful aspects of
the extension.  My own interest is aimed more at embedding metadata
about the nested resource than embedding the resource itself.

thanks-
peter keane

What Google's feedLink extension provides, that isn't currently achievable
with a standard atom:link element, is the ability to embed the linked
resource itself inside the entry. Such functionality has many much more
applications than just embedding related feeds though. Why not embed a
related image, or a related html page?

Either way, whether you want to link to a related feed or you want to embed
a related feed, the feedLink extension seems to me a pretty lousy way to go
about it.

Regards
James






Reply via email to