- From: Mark Baker <distobj@acm.org>
- Date: Fri, 13 Aug 2004 11:00:47 -0400
- To: hugo@w3.org
- Cc: www-archive@w3.org
Hi again Hugo,
I'm not sure why I didn't just point this out in the first place, but
there's a fine example of the EPR/URI issue in the submission itself;
<wsa:EndpointReference xmlns:wsa="..." xmlns:fabrikam="...">
<wsa:Address>http://www.fabrikam123.example/acct</wsa:Address>
<wsa:ReferenceProperties>
<fabrikam:CustomerKey>123456789</fabrikam:CustomerKey>
</wsa:ReferenceProperties>
<wsa:ReferenceParameters>
<fabrikam:ShoppingCart>ABCDEFG</fabrikam:ShoppingCart>
</wsa:ReferenceParameters>
</wsa:EndpointReference>
How is that not a URI like this?
http://www.fabrikam123.example/acct/123456789?ShoppingCart=ABCDEFG
This seems to me to be a *major* problem with WS-Addressing with respect
to Web architecture.
Now, arguably, since the spec seems to be ambiguous about whether or not
visibility into the reference properties/parameters is required by
consumers of the EPR, it may or may not be necessary to peek inside an
http EPR URI - which would of course be a violation of URI opacity "good
practice"[1]. But worst case, EPRs should use a URI scheme which
licenses the kind of peeking that WS-Addressing specifies (even though
I don't think WS-Addressing needs to do that, as they could be fully
opaque and still do their job as identifiers, since it seems that only
the service has to actually process them).
[1] http://www.w3.org/TR/2004/WD-webarch-20040705/#uri-opacity
Cheers.
P.S. I'm on vacation for the next week, and will be unreachable during
that time.
Mark.
--
Mark Baker. Ottawa, Ontario, CANADA. http://www.markbaker.ca
Received on Friday, 13 August 2004 14:59:46 UTC