[Standards-JIG] Proposed Changes to JEP-60 - PubSub
Fletcher, Boyd C. J9C534
Boyd.Fletcher at je.jfcom.mil
Thu Jun 3 21:16:44 UTC 2004
> -----Original Message-----
> From: standards-jig-bounces at jabber.org
> [mailto:standards-jig-bounces at jabber.org] On Behalf Of Richard Dobson
> I think I can see why you are getting so confused, JEP-0060
> is supposed to be just the building block to build full
> featured pubsub applications, it doesnt really seem that this
> JEP is really intended to be used to create a pubsub server,
> as you say it leaves lots of things as optional which is fine
> as it is a building block, to be useful for building a pubsub
> server it might be better if we specify it in its own JEP
> rather than altering this one a unnecessarly restricting
> future pubsub applications ontop of this, either we need a
> new JEP based on JEP-0060 that locks down most of the
> optionals saying all the options MUST be supported in a
> pubsub server or pubsub server implementation notes need to
> be added to a similar effect without altering the rest of the JEP.
ya'll missing the point. The current JEP is not sufficiently detailed enough to build anything but trivial PubSub applications. the changes I am recommending are not limiting and expland and enhance the pubsub specification so that it can be used as the lower level protocol for an unlimited number and type of PubSub applications. For some reason, many of ya'll believe that tigtening up a specification makes it less flexible, while reality it's the exact opposite. Concise, non-ambiguous, feature complete specifications entice vendors and developers to want to implement the protocol because that can reasonbly expect that anyone who is compliant with the specification will have a certain level of functionality. I have spoken with several of the jabber servers vendors about this and the reason they have been reluctant to implement many of the JEPs is: they constantly changing and they have a high degree of ambiguity.
> Well you did say the following and I quote:
> > We are not limiting the functionality of the base protocol. We are
> > significantly enhancing it by specifying how it behaves in a
> > predictable, well understood manner. So that clients can implement
> > User Moods, Headline news, avatars, etc.... WITHOUT having
> to have a
> > specific custom PubSub implementation on EVERY server you
> have an account on.
> This seems to suggest that you think that a pubsub server
> based on JEP-0060 needs to understand the higher level
> protocols to work, when it clearly doesnt if you really are
> understanding what parts of this relate to implementing a
> pubsub server.
absolutely not. I'm saying the that core protocol should be sufficiently robust enough so that a SINGLE pubsub server can be written and the high level protocol is handled completely by the jabber client.
> Also in what way does this protocol stop you from being able
> to create higher level protocols based on it? I cant see any
> real reasons, the only real reasons you have presented all
> have to do with implementing a pubsub server, not higher
> level protocols, infact others dont seem to have had any
> problems creating protocols that work based on JEP-0060 (User
> moods for example), so your argument that you cant build
> protocols does not seem valid at all as there is lots of
> evidence to the contrary.
the current batch of PubSub "applications" are so trivial, they don't run into the problems I describing, but try implementing something like Whiteboarding (like we are) and you start running into lots of issues with the inadequacies of the spec.
More information about the Standards