- From: Monica J. Martin <Monica.Martin@Sun.COM>
- Date: Wed, 09 May 2007 12:20:19 -0700
- To: Felix Sasaki <fsasaki@w3.org>, Sergey Beryozkin <sergey.beryozkin@iona.com>
- Cc: "'David Orchard'" <dorchard@bea.com>, public-ws-policy@w3.org
> Sergey Beryozkin wrote: Hi > Do you think wsp:optional in <assertionWithNoWireEffects optional="true"> > has any semantics, where assertionWithNoWireEffects means that the client > does not need to do anything on the wire even if it recognizes this > assertion ? > The reason wsp:ignorable is here is because the above is considered a > "lie". > Therefore wsp:optional=true has semantics. My point is that to make > wsp:ignorable truly useful and unambiguous the combination of > wsp:optional=true and wsp:ignorable=true should be prohibited. sasaki: I can't estimate the impact of prohibiting a combination of two units, where one (wsp:Optional) does not exist in the policy data model. But since you wrote "mabye in v.Next", you seem to agree that such considerations need more time and are not on our plate currently? >>monica: What this alludes to (given Sergey's query and your response) a question I have posed to Dave Orchard: Does the use of Ignorable transcend the non-wire boundary of which we defined it particularly considering the Versioning example of whcih we agreed? Sergey, I think you could be saying, we shouldn't traverse that boundary. And, if we do allow that boundary to be traversed, what should we do about it now or later? Thanks.
Received on Wednesday, 9 May 2007 19:19:41 UTC