[standards-jig] Publish-Subscribe JEP status

Jeremie jeremie at jabber.org
Thu Apr 11 18:19:03 UTC 2002


> It seems to me that the frustration comes when all of a sudden JINC
> comes in with a new idea that they never mentioned they are working on. 
> I can understand this from a commercial R&D perspective, but when there
> is a public topic such as pubsub, that has been public for a long time
> it's very easy to understand the frustration.  It purely appears that
> the community input is being disregarded and that JINC will take their
> own course of action no matter what, because that's how they
> investigated it.  I honestly have no idea about how to combat this, but
> I hope I can help more completely identify the problem.

I believe the misunderstanding here is simple.  From what I understand
(and learned this week when I was in Denver), JINC has been working on a
pub/sub protocol but not in the same context as what everyone else here is
talking about.  Their angle is purely behind the scenes, something that
components inside their server use to communicate, analogous to how XDB is
used in jabberd-1.4.  They were not developing it with the focus or
intention of it being an iq or exposed protocol as others have been.

So, in their process of defining that use of pub/sub, they've learned some
things and found some similarities that they want to share and that might
help define a more complete overall Jabber pub/sub protocol, which is a
good thing :)

I suspect that most of the confusion on this is just some poor
communication, they've clearly been doing some thing different and I don't
think they saw the direct overlap or potential for sharing until they got
to a point recently where they really defined what they needed for their
internal-component pub/sub.  Moving forward though it's obvious all the
approaches share some things in common, and all parties can work together
to iron out a solution for as many of the needs as possible.

Jer




More information about the Standards mailing list