On May 10, 2006, at 12:11 PM, Joe Gregorio wrote:
On 5/10/06, John Panzer <[EMAIL PROTECTED]> wrote:
Does this imply that a service cannot also do content negotiation
if no
preference is specified in the URI (if a "base URI" alone is
used)? I don't
know.
There is a separate question lurking here: Whether or not we
mandate
multiple URIs, or allow them, or ban content negotiation... is
there one
resource hiding behind the URI(s) or two?
I responded to Bill on this earlier in this thread:
(http://www.imc.org/atom-protocol/mail-archive/msg05092.html)
"Does it matter? There are two URIs that have different
representations, whether or not those are the 'same' resource
is a server implementation detail and as far as I can tell
has no impact on the interop of the protocol."
In the long run, yes, as features get added in standards or in
implementations, it really matters.
- Do the two URIs have the same behavior in terms of access
control, and show the same ACL
- Would locking one cause the other to be locked
- Do they have the same metadata, e.g. creation date and last-
modified date, owner, etc.
- Do they both get deleted together
(Reading Tim's very different answer, he's right about the general
case talking to any old server, but he might be wrong if a specific
standard says "There is one resource behind these two URIs" and you
know you're talking to a server that supports that specific standard.)
Lisa
