[Standards-JIG] Re: JEP-0163: further PEP simplification

[Ian Paterson]

I agree this is an *inspired* change to the protocol.

>> What about cases where users want to subscribe to an "Open" access  model 
>> PEP node, but they are not subscribed to the owner's presence?
>
> I have always preferred limiting it all to presence-based publishing.  I 
> don't see a problem with requiring subscription, it simplifies a lot.

To fit with existing JEPs based on JEP-0163 (and many future use cases), the 
owner must still be able to specify the "Open" access model. Of course this 
proposal means non-presence-subscribers will not be able to subscribe to 
pubsub nodes, but they will at least still be able to use the get-items 
protocol.

There are real consequences of this restriction. Some PEP data will be 
public, won't change often, but may be used frequently. e.g. public-keys (in 
JEP-0189), e.g. vCard-style profile data... So we may see plenty of 
'unnecessary' get-items requests to check if the published data has changed.

Even so, I can live without JEP-0060-style subscriptions for the 
non-Presence access models in JEP-0163 (the subscription protocol could be 
retrofitted quite easily for these cases if we decide we need it).

- Ian