- From: Yaron Goland <yarong@Exchange.Microsoft.com>
- Date: Mon, 3 Jan 2000 10:35:46 -0800
- To: "'Geoffrey M. Clemm'" <geoffrey.clemm@rational.com>, w3c-dist-auth@w3.org
Geoff said: "It sometimes acts like a resource (modifies the parent collection), and sometimes does not (returns a 404 when you GET it)." In HTTP, as I understand it, one sends a method to a resource identified by a URI. As such any response one gets must have come from a resource. Therefore a 404 must be issued by a resource. This is why the object model I suggested for WebDAV has the universal NULL resource that handles things such as issuing 404s. As such how can one say that because a resource issues a 404 it is not acting like a resource? Yaron > -----Original Message----- > From: Geoffrey M. Clemm [mailto:geoffrey.clemm@rational.com] > Sent: Thursday, December 30, 1999 7:35 PM > To: w3c-dist-auth@w3.org > Subject: Re: Creating a lock-null in a locked collection > > > > From: Joe Orton <joe@orton.demon.co.uk> > > ... It just seemed slightly weird, that if you LOCK or UNLOCK a > lock-null resource, you are modifying the state of the > parent collection; > when if you do the same to a normal resource, you are not. > > Yes, that is one of the reasons I strongly object to the notion of a > "lock null resource". It sometimes acts like a resource (modifies the > parent collection), and sometimes does not (returns a 404 when you GET > it). This makes it very hard to predict what its behavior should be > whenever you extend the protocol (i.e. should it act like a resource, > or act like not a resource). For example: > > - the BIND protocol (can you "BIND" a lock null resource to > another URL?) > - the versioning protocol (do you have to checkout a > versioned collection > in order to add a lock null resource to it?) > > Cheers, > Geoff >
Received on Monday, 3 January 2000 13:36:45 UTC