Re: [VCARDDAV] Vcarddav: creating new resources

[Julian Reschke <[e-mail]>]

Hi,

Cyrus encouraged me to bring up this topic again.

Last month's discussion ended around 
<http://www.ietf.org/mail-archive/web/vcarddav/current/msg00650.html>, 
with me reporting back from the related discussion over at the HTTPbis 
WG's mailing list 
(<http://lists.w3.org/Archives/Public/ietf-http-wg/2008OctDec/0096.html>).

At this point I think this WG needs to decide whether the "server picks 
the URI" approach should be *allowed*.

If it's not, the carddav specification clearly should state this, and in 
particular rule out those workarounds we're seeing in CalDAV (which got 
discussed over here and on the HTTPbis mailing list).

If it should be allowed (and I think it should), we should add a way to 
do it. PUT does not work for the reasons that came up during the 
discussion (please read the discussion on the httpbis WG mailing list). 
My proposal would be to use "Using POST to add Members to Web 
Distributed Authoring and Versioning (WebDAV) Collections" 
(<http://greenbytes.de/tech/webdav/draft-reschke-webdav-post-latest.html>) 
-- an alternative would be a new verb as proposed earlier ("ADDMEMBER").

If this WG wants to use this proposal, I'm happy to hand it over to the 
WG for completion (I think there are few things left to do, mainly 
taking out the stuff that relates to the unfinished HTTP Link header 
draft, and potentially moving the elements into the "DAV:" namespace).

BR, Julian



Julian Reschke wrote:
> Julian Reschke wrote:
>> ...
>> POST is not a new method.
>>
>> PUT+Location definitively does *not* work everywhere; I'm sure of that 
>> because the WebDAV client code I've written in the past certainly 
>> won't understand it.
>>
>> If the catch-all semantics of POST are a problem, then one way to 
>> handle  this is to offer the client enough information about what POST 
>> semantics are implemented. For instance, AtomPub does this by 
>> requiring the more specific behavior for the Atom feed URIs it 
>> exposes. My proposal (for WebDAV) does the equivalent; it exposes the 
>> URI for "add-member" POSTs on the wire (PROPFIND or DAV:error), so a 
>> client can rely on the desired semantics.
>>
>> BR, Julian
>> ...
> 
> For the record, we *did* move the discussion over to the httpbis WG's 
> mailing list; see thread at 
> <http://lists.w3.org/Archives/Public/ietf-http-wg/2008OctDec/thread.html#msg44> 
> in general, and Roy's message at 
> <http://lists.w3.org/Archives/Public/ietf-http-wg/2008OctDec/0096.html> 
> in particular.
> 
> BR, Julian


_______________________________________________
VCARDDAV mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/vcarddav