[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Issue verification: 2.23, 2.36, 2.39
Verified against base draft j; I have tiny quibbles with 2.23 & 2.39. 2.23: Is subscription a WS-Resource? Approach taken as agreed, but... Lines 469-472: Unequivocally YES Lines 1106-1114 (and many others): Support for optionality, as specified in the proposed recommendation. The text from the proposed recommendation has been incorporated, so the dependency on WS-RL is gone. However, lines 469-472 state with absolutely no ambiguity that a subscription reference is always to a WS-Resource. I think this is as intended, however, as a non-participant in WS-RF, I find it somewhat confusing. I interpret this to mean that the reference has the appropriate form for a WS-Resource, but no implied WS-Resource semantics? 2.36: Specify baseline Policy behaviour: OK Subscription durability (Pause/Resume) line 1292-1294: OK Message durability: No relevant text in Base: OK Pull Notification: OK lines 868-1104 Since PullPoint is a separate interface now, AFAICT, there is no need for a baseline policy; either you know where there are pull point creators that can create pullpoints that are reachable by both your notification producer and your consumer, or not, and the notification producer doesn't know the difference anyway. Message continuity (Lines 1156 - 1464): OK Flow control (lines 630-631): OK No normative text associated with this, however, I the use of it as an example in the SubscriptionPolicy element description provides adequate guidance to implementors that there is no guaranteed flow control. Reliability (lines 277-283): OK In the new introductory text for "production vs. delivery", I think we imply quite strongly that reliable delivery is not considered to be in scope of the proposal. I disagree with Sanjay's earlier assertion in [1] that this should be covered in the NotificationConsumer.Notify() section, since this section emphasizes the responsibilities of the consumer, not the producer or the delivery channel. 2.39: Tighten up the semantics of "send/delivery": Done well enough, but... OK uses of "send": In the context of "send a fault": OK Non-normative text associated with motivation for the pull interface: Questionable, but OK In the context of "send a WS request": OK In the context of "send a comment to": OK Line 141: "send information to": OK I'd change these: Line 521: Should be "NotificationProducer should produce messages for the NotificationConsumer". Line 1531: Should be "refuse to *produce* messages to Notification Consumers that are not known to be authorized." The term "deliver" is not used anywhere in the document. The term "delivery" is used exactly as intended in all cases. [1] http://www.oasis-open.org/apps/org/workgroup/wsn/email/archives/200504/m sg00023.html
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]