RE: proposal for 7478

I think the current spec is pretty clear that there are two ways a 
subscription can end.  One is due to hitting the expires time and the 
other is due to "some other reason".  No one, I think, will dispute that 
"some other reason" will always be possible. However, the expires time is 
all about the normal termination of a subscription and that needs to be 
very well understood and agreed to.  Think of it this way, a normal 
expires will NOT generate a SubscriptionEnd message, while "some other 
reason" will (assuming an EndTo is provided).  The expires time 
negotiation is all about the times when the SubscriptionEnd message is not 
involved.  These two situations are not the same as you are implying - 
they have very different behavior - and (Gil should like this) the state 
tables should show this difference.  :-)

thanks
-Doug
______________________________________________________
STSM |  Standards Architect  |  IBM Software Group
(919) 254-6905  |  IBM 444-6905  |  dug@us.ibm.com
The more I'm around some people, the more I like my dog.



"Li, Li (Li)" <lli5@avaya.com> 
09/17/2009 10:06 AM

To
Doug Davis/Raleigh/IBM@IBMUS
cc
<public-ws-resource-access@w3.org>, 
<public-ws-resource-access-request@w3.org>
Subject
RE: proposal for 7478






Doug,
 
I will focus on my response only to your part 2) as my previous email did. 

 
What concerns me is the semantics of ?minimum? expiry time. If a 
subscriber asks for an expiry time 1 hour and tags it as minimum, can the 
event source terminate the subscription before 1 hour? According to 
current spec, it can. And this is correct, because no event source can 
predict the future and guarantee that the subscription will last for 1 
hour. Therefore, a subscription with a minimum 1 hour expiry is the same 
as a subscription with a ?random? 1 hour expiry (one created without the 
minimum tag). Since there is no difference in actual lifetime between two 
subscriptions, adding such a tag has no consequence but creates confusions 
for the subscribers. 
Thanks,
Li Li  

From: Doug Davis [mailto:dug@us.ibm.com] 
Sent: Wednesday, September 16, 2009 11:52 AM
To: Li, Li (Li)
Cc: public-ws-resource-access@w3.org; 
public-ws-resource-access-request@w3.org
Subject: Re: proposal for 7478
 

Li wrote: 
...
> By change 1), it assumes all subscribers will never accept any
> subscriptions with a shorter expiration. Therefore, there is no need to
> create such subscription in the first place. While this may be the case
> in some situations, I'm not sure WS-E should enforce this assumption for
> all situations. In some situations, a subscriber may want to accept a
> subscription with a shorter expiration, with the hope that it can renew
> it later on when the event source is less loaded. 

this "hope" part is what worries me. W/o a clear guarantee that the 
subscription 
can be renewed then the subscriber is taking a large gamble.  It seems to 
me that 
in order to ensure we have a properly interoperable spec we need to be 
very 
clear about what both sides expect and want.  Gil's proposal does this but 

you are correct that it assumes the subscriber only cares about the 
minimum duration. 
You point out that a subscriber might really be ok with the uncertainty 
around being able to renew (the 'hope' part) - but if so then it needs to 
tell the source this bit of information. 

To me there are two parts of this: 
1 - we need to make sure that the source advertises what it supports by 
using policy.  This means that if we keep duration and dateTime then there 
is 
no interop problem as long as the source tells people which it supports. 

2 - we need to make sure that the subscriber tells the source what it 
expects w.r.t. the new subscription.  This means that when it asks for an 
expires time it needs to not only tell it the duration/dateTime, but it 
should also indicate whether this is an upper limit or a lower limit, or 
even just a suggestion.  Perhaps a new attribute on the Expires element to 

indicate this would do it.  W/o this flag I don't think we can get the 
level of interop we want by sticking with the current "random" expires 
time 
approach. 

thanks, 
-Doug