Re: Spam: Re: [ws-caf] policy ai

[Mark Little <mark.little@arjuna.com>]

Green, Alastair J. wrote:

>I took the phrase "reasoning about the relationships" to be of wider
>significance -- similar language has been used in your and Greg's
>discourse on the (alleged) need to know what the endpoints are up to,
>and thence to the idea of an ACID-only protocol, which I think is a
>rather stunted beast. 
>
>The fact that the declarative use of the transaction context on receipt
>by a bean is a private matter is one example of the ability to be
>largely ignorant of what transactions means to the interlocutor, but
>still be able to make use of them.
>
>On the narrow point, it seems that any way of marking an endpoint as
>being capable of receiving a context is better than none. 
>
>In that respect I think that fixing on something that actually works is
>better than trying to second-guess the next three twists and turns on
>the secondary roads running through WS-Land. 
>
>Is there a way of doing this which will work with current WSDL and not
>just 2.0?
>  
>
Well, assuming we were to adopt WS-Addressing, one use for the MetaData 
bucket is policy (at least it was prior to WS-Addressing going into W3C, 
but has since been changed). Check out Steve's article: 
http://www.iona.com/hyplan/vinoski/pdfs/IEEE-WS-Addressing_Metadata.pdf

I believe that tying policies to the EPR is better than the WSDL. It is 
like comparing transactionality via IOR attribute "bit" versus 
transactionality via inheritence in the OTS.

Mark.

>Alastair
>
>Alastair J. Green
>CEO and CTO
>Choreology Ltd
>68 Lombard Street
>London EC3V 9LJ
>www.choreology.com
>
>+44 870 739 0050
>+44 870 739 0051 (fax)
>+44 795 841 2107 (mobile)
>
>
>-----Original Message-----
>From: Mark Little [mailto:mark.little@arjuna.com] 
>Sent: 26 July 2005 15:31
>To: Green, Alastair J.
>Cc: Greg Pavlik; ws-caf
>Subject: Spam: Re: [ws-caf] policy ai
>
>
>
>Green, Alastair J. wrote:
>
>  
>
>>AJG: This is, I think, a slight overstatement :-). You can reason, but
>>only within limits -- and without requiring omniscience. 
>>
>>[stuff deleted]
>>
>>    
>>
>I think Greg's point was more that a user of an EJB cannot tell what the
>
>transactional requirements are for the EJB. There's no equivalent of 
>CosTransactions::TransactionalObject and because IIOP is not mandated, 
>there's no way of delving within the object's IOR at the client side.
>
>Mark.
>
>  
>

-- 
Mark Little
Chief Architect
Arjuna Technologies Ltd
www.arjuna.com