On 13.04.2010 01:49, Stephen Pendleton wrote: > -----Original Message----- > From: [email protected] [mailto:[email protected]] On > Behalf Of Florian Zeitz > Sent: Monday, April 12, 2010 6:58 PM > To: [email protected] > Subject: Re: [Standards] XEP-0080 interoperability > > On 12.04.2010 23:45, Stephen Pendleton wrote: > >> Maybe my view on this diverges from everyone else's, but I don't think >> PEP is/should be restricted to IM. It is a totally artificial >> restriction you are putting on your self that makes things more >> complicated. Why wouldn't (to keep the example) one microblogger be >> connected to one JID and his microblog stored in a PEP node on that? > > I don't think PEP is restricted to IM I guess, but it does require a > roster/presence to make it work. In fact, I have done this by creating a web > service that uses a bot to receive PEP events (location, etc) from users > subscribed to the bot. If you had 1000 entities who wanted to get updates to > your microblog, you would need 1000 entities subscribed to you in your > roster in order for PEP to work. > > I have even a better example: I want to post your current location to a web > page by subscribing to your geoloc pubsub node. Currently I would have no > way to programmatically find out where your node is I believe. >
I'm by no means a PubSub or PEP expert, so maybe I'm wrong. But AFAICT the XEP clearly states that a PEP service MUST support the "open" access model as well as the subscribe use case. It also states "An entity shall receive event notifications if [...] The node has an open access model and the entity has explicitly or implicitly subscribed to the node as explained in XEP-0060." Wouldn't that suffice?
