[standards-jig] Pub/Sub Explanation

Adam Theo theo at theoretic.com
Tue Apr 16 18:23:29 UTC 2002


Dave Smith wrote:
> Greetings...
> 
> Wow. There are some fundamental differences in our trains of thought.
> 
> Frankly, what I'm hearing from DJ and piers is the requirement for a
> light-weight notification system -- and maybe that's a truer definition of
> "pub/sub". What I am proposing is a system which is a super-set of the
> functionality there...a sort of "persistent pub/sub". There's certainly no
> reason why a system such as I described can't implement the functionality
> described in JEP 24.
> 
> With this in mind, perhaps we can agree that these are two separate systems
> and as such should be pursued independently of each other.

No, I don't think that is needed at all. How about we tackle both trains 
of thought, but treat DJ's as the "core protocol for the Ultimate 
Generic PubSub", separate from any storage or management functionality. 
Then we take a storage "add-on" or "extension" that could prossibly be 
independant in its own right (not inherently tied to pubsub) and a 
topic/thread management mechanism (also not inherently tied to pubsub) 
and tie the three together for Dave's work? Much of this storage and 
topic management would be based on JINC's work since they are much 
farther along in that, I gather. The key thing is to keep everything 
modular, and most functionalities separate from the others so they can 
be used independantly.

I'm just saying don't give up hope yet that a universal pubsub can be 
made that makes everyone happy.

-- 
     /\  Adam Theo, Age 22, Tallahassee FL USA
    //\\   Email & Jabber: theo at theoretic.com
   //  \\  (Boycotting AOL, therefore no AIM or ICQ)
=//====\\=  Theoretic Solutions: http://www.theoretic.com
//  ||  \\     "Bringing Ideas Together"
     ||      Jabber Protocol: http://www.jabber.org
     ||         "The Coolest IM on the Planet"
     ||  "A Free-Market Socialist Patriotic American
     ||      Buddhist Political Philosopher."




More information about the Standards mailing list