http://stage.vambenepe.com/archives/43 IT management in a changing IT world Thu, 20 Nov 2008 11:41:52 +0000 http://wordpress.org/?v=2.6.3 http://stage.vambenepe.com/archives/43#comment-8 vbp Thu, 30 Nov 2006 02:45:32 +0000 http://stage.vambenepe.com/archives/43#comment-8 Hi Mark, sorry for the delay in replying. I agree with you and I am not suggesting that this feature be necessarily added to the base WS-Addressing spec. I'd like WS-Addressing to fix the @isReferenceParameter mechanism, but the feature I describe can very well exist as a separate spec that only comes into play in the (mostly model-driven) scenarios in which it is useful. Hi Mark, sorry for the delay in replying. I agree with you and I am not suggesting that this feature be necessarily added to the base WS-Addressing spec. I’d like WS-Addressing to fix the @isReferenceParameter mechanism, but the feature I describe can very well exist as a separate spec that only comes into play in the (mostly model-driven) scenarios in which it is useful.

]]> http://stage.vambenepe.com/archives/43#comment-7 marklittle2005 Thu, 30 Nov 2006 02:44:31 +0000 http://stage.vambenepe.com/archives/43#comment-7 Hi William. I agree that the use case you've mentioned is interesting, but I don't see why this has to be done within the addressing infrastructure. This kind of QoS-related endpoint referencing has been done before in several distributed environments and is either handled directly at the application level or within some intermediate trading service. I don't see a compelling reason to put this kind of functionality within the base WS-Addressing specification, which is complex enough as it is. Apart from the fact I believe it can be better handled higher up the stack, it's also likely that such functionality won't be a requirement of the majority of WS-Addressing use-cases, so why clutter the foundation? Hi William. I agree that the use case you’ve mentioned is interesting, but I don’t see why this has to be done within the addressing infrastructure. This kind of QoS-related endpoint referencing has been done before in several distributed environments and is either handled directly at the application level or within some intermediate trading service. I don’t see a compelling reason to put this kind of functionality within the base WS-Addressing specification, which is complex enough as it is. Apart from the fact I believe it can be better handled higher up the stack, it’s also likely that such functionality won’t be a requirement of the majority of WS-Addressing use-cases, so why clutter the foundation?

]]>