W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > April to June 2005

Re: WGLC draft-ietf-webdav-redirectref-protocol-12

From: Lisa Dusseault <lisa@osafoundation.org>
Date: Tue, 14 Jun 2005 10:48:32 -0700
Message-ID: <42AF1870.5030000@osafoundation.org>
To: Julian Reschke <julian.reschke@gmx.de>
Cc: Cullen Jennings <fluffy@cisco.com>, WebDav <w3c-dist-auth@w3.org>, Jim Whitehead <ejw@soe.ucsc.edu>, Geoffrey M Clemm <geoffrey.clemm@us.ibm.com>

I think that hiding redirects from authoring clients may be a perfectly 
good and reasonably interoperable thing to do.  It depends on two 
things: the model and the need.

The model that REDIRECT has is very pure -- that a URL that produces a 
redirect is a resource.  That's not the only model, however.  An 
implementor (client or server) could have a model where a URL that 
produces a redirect does *not* point to a resource, just like a URL that 
produces a 404 does not have an associated resource.  That model is 
already reflected in software like mod_rewrite, a module which is quite 
incompatible with REDIRECT.  The "redirect is not a resource" model 
could still support authorability, for example by adding properties to 
collections, and the properties could contain redirect rules. 

The "need" is the practical concern which we've already discussed, but 
I'd add that if we have a way to author redirects that is incompatible 
with mod_rewrite and the way administrators actually need to create 
redirect rules, then we may not have even solved the theoretical need.

The fact that SAP (aka Julian's org) implements them as resources is in 
fact very suggestive.  If that's the model you thought of, perhaps it's 
a better model for interoperability, a more WebDAV-suited model.  So now 
we're back to wanting other implementations to validate that model.


Julian Reschke wrote:

> Lisa Dusseault wrote:
>> The main flaw in the REDIRECT proposal is that there is not enough in 
>> the way of plans to implement it.  Without a set of independent 
>> implementors around to review it, I fear it's too complex or has 
>> missed key interoperability issues.
>> ...
> Forgot to mention: actually it *adresses* interoperability issues, 
> because redirect-type resources already exist on the Web, but RFC2518 
> fails to describe how to handle them (at least) in PROPFIND responses.
> Authorability may not be as important, but discovery certainly is. 
> Today, Apache/moddav just hides redirects from WebDAV clients. I think 
> this is a problem that needs to be solved.
> Best regards, Julian
Received on Tuesday, 14 June 2005 17:48:40 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:01:34 UTC