[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Simple] WGLC: XCAP Base



inline.

Aki Niemi wrote:

I don't know. I fear that the answer is no; after a DELETE, the resource is deleted, so by definition THAT resource doesn't have an etag anymore.


There is something with Etags here that keeps worrying me. Because of
the way that etags are distributed over an XCAP document (i.e., all
elements/resources share the same etag value), it seems that we lose
some important property for some XCAP applications.

Let's say we have an XML document that has two sub trees. In one subtree
(called <bar>) you have a flat list of properly indexed, simple
elements, like


<foo name="x">TRUE</foo>

Naturally updating this list doesn't require safe updates at all. One
could simply PUT/GET/DELETE blindly using the name attribute as
selection key.

Now let's say that the other subtree (called <baz>) is a much more
elaborate object, where safe updates are required. SOmething akin to
common policy <rules> for example.

Because of the way we have currently defined etags usage, I believe any
blind update to the <bar> subtree automatically forces also getting the
<baz> subtree, because the etags have changed.

I think this is really suboptimal, and it would be great if it could be
somehow avoided. For example, having the possibility for an application
to define etag "ranges". In the above example, there would be three
ranges defined: One for the root element of the document, second for the
bar subtree and the third for the baz tree. A change would only
propagate within a single range.

We've been through this, Aki.

I proposed extensive text onto the list (on 6/24/04) on a framework for Etag scopes that we *could* support in order to allow what you are talking about. As I wrote up the details, the complexity just grew and grew, and it hit the point that I didnt have any good ideas on how to practically implement it. As such, an open issue was raised (Issue 8), and the list consensus was to simplify, and keep etags scoped to a document.

We also heard from Scott Lawrence today that having the etag scope to a document was really an essential feature for using HTTP as well.

As such, I am disinclined to revisit this open issue, as we've been through it a few times already. Unless you can bring a new workable proposal to the table, I think we should stay with the current solution.

-Jonathan R.



--
Jonathan D. Rosenberg, Ph.D.                600 Lanidex Plaza
Chief Technology Officer                    Parsippany, NJ 07054-2711
dynamicsoft
jdrosen at dynamicsoft.com                     FAX:   (973) 952-5050
http://www.jdrosen.net                      PHONE: (973) 952-5000
http://www.dynamicsoft.com

_______________________________________________
Simple mailing list
Simple at ietf.org
https://www1.ietf.org/mailman/listinfo/simple