So we have had this discussion several times now. I suggest that when the WS-Addressing 1.0 WSDL Binding becomes a recommendation that the wsa:Address will not be treated by the tooling as a URL. Further the fact that current tooling gets it wrong is not surprising given that there is no normative profile for WS-Addressing. While I understand the tooling issues I would be extremely hesitant to base an architectural decision on vagaries of tooling. Let me be very clear about this an EPR does not have a 1 to 1 correspondence with a protocol and binding address. It MUST be true that an EPR can support multiple protocols and their associated addresses. With that as a base understanding is there any argument, other than URL tooling, that you can articulate against using the wsa:Address IRI field as a unique resource identifier? Tom -----Original Message----- From: owner-ogsa-wg@ggf.org [mailto:owner-ogsa-wg@ggf.org] On Behalf Of Mark Morgan Sent: Tuesday, November 15, 2005 11:11 AM To: 'Steve Loughran' Cc: ogsa-naming-wg@ggf.org; ogsa-wg@ggf.org Subject: RE: [ogsa-wg] Re: [ogsa-naming-wg] ... without wsa:Address profile, AbstractName is meaningless ... I'm concerned that by saying that we will start using the Address field for uniqueness and comparison reasons that that is going to place a MUCH heavier burden on the tooling, programers, and users then the abstract name does currently. All of my web services right now have URLs in the Address field because that's what the tooling supports and clearly the URL is not unique for all resource endpoints (I in fact have thousands of resource endpoints residing on my machine right now at exactly one URL). Using IRIs or URIs in the address field of the EPR is great on paper, but I don't know of any tooling that in fact supports that.
-----Original Message----- From: owner-ogsa-wg@ggf.org [mailto:owner-ogsa-wg@ggf.org] On Behalf Of Steve Loughran Sent: Monday, November 14, 2005 12:16 PM Cc: ogsa-naming-wg@ggf.org; ogsa-wg@ggf.org Subject: Re: [ogsa-wg] Re: [ogsa-naming-wg] ... without wsa:Address profile, AbstractName is meaningless ...
Hi Frank, The Web Architecture (http://www.w3.org/TR/webarch/#identification) discusses and provides guidance on the issues you describe of IRI collision, IRI aliases etc. Does this document not address
Mark McKeown wrote: the issues
of the profile work you are suggesting?
When you advocate using the wsa:Address for the AbstractName are you saying that there is no need for adding the AbstractName element from WS-Naming into the EPR or are you saying that the wsa:Address (or some combination of the wsa:Address+wsa:ReferenceParameters) should be used in the AbstractName element.
eg. <wsa:EndpointReference xmlns:wsa="http://www.w3.org/2005/03/addressing" xmlns:name="http://ggf.org/name"> <wsa:Address>http://ggf.org/example/B944388</wsa:Address>
<name:AbstractName>http://ggf.org/example/B944388</name:AbstractName>
</wsa:EndpointReference>
And if so, is everyone expected to (a) have read and understood section 3.2.3. on the subject of URI equivalence(below), or (b) to explicitly require case-sensitive-string-matching as the comparison.
In the case above, are the following abstract names equivalent?
<name:AbstractName>http://GGF.ORG/example/B944388</name:AbstractName> <name:AbstractName>http://GGF.ORG:80/example/B944388</name:Abs tractName> <name:AbstractName>http://ggf.org:/example/B944388</name:AbstractName> <name:AbstractName>HTTP://ggf.org/%65xample/B944388</name:Abst ractName>
I ask, as for the test document for CDDLM, I need to define equivalence of things. And as EPR equivalence is so troublesome I am looking at WSDM ResourceIDs, but even there you have to specify what equivalence logic you will be using.
-steve
--------------------- RFC2616 on comparisons. Note that java.net does not implement this logic:
3.2.3 URI Comparison
When comparing two URIs to decide if they match or not, a client SHOULD use a case-sensitive octet-by-octet comparison of the entire URIs, with these exceptions:
- A port that is empty or not given is equivalent to the default port for that URI-reference;
- Comparisons of host names MUST be case-insensitive;
- Comparisons of scheme names MUST be case-insensitive;
- An empty abs_path is equivalent to an abs_path of "/".
Characters other than those in the "reserved" and "unsafe" sets (see RFC 2396 [42]) are equivalent to their ""%" HEX HEX" encoding.
For example, the following three URIs are equivalent:
http://abc.com:80/~smith/home.html http://ABC.com/%7Esmith/home.html http://ABC.com:/%7esmith/home.html