Thanks for the quick reply .
Is the source already available somewhere? Is there a demo installation?
We have been using the system for about a year or two now in production. The source and the demo can be made available but are currently not.
My query is an independent, preliminary investigation to see what kind of things we might need to do to prepare for the first open source release. I will have to check with my academic department before uploading code or releasing a URL to a demo server. An e-mail or message thread might be more appropriate for a detailed conversation.
However, to give you an idea of installation, essentially:
-
Setting up Openfire + BOSH + Strophe | ExpertNotFound
-
Then Tomcat + WebApplication war file
In order to give you an idea of what the web application is about, here is a screenshot from my localhost:
Does the project depend on Openfire? Or was it build on top of Openfire’s fastpath plugin?
The project was built to be independent of Openfire. It is not built on top of the fastpath plugin.
We did, however, rely heavily on Smack (Jabber client) which may have one or two quirks when used with different systems (I haven’t tested with different chat servers yet).
How’s it different from XEP-0142: Workgroup Queues? … But ideally it’s designed as XMPP component which can be used with any XMPP server supporting external components.
So, if I remember correctly, the issue was small but significant. And perhaps with slight modification to our system, we could fit into that paradigm.
So, I wanted to use FastPath, but the big issue brought up by my client was that the Service in the Workgroup Queues do not broadcast requests from the Users. Instead, in the FastPath implementation, the Service cycles through all available Users and each User must “accept or reject chat requests”.
This causes delay time and confusion, especially when some of our Agents might be busy or unresponsive.
All of the other aspects of the web application use XMPP components. We automatically set up MUCs as our incoming tickets. The agents wait in an agent MUC until they claim an incoming ticket (aka join a MUC with a User).
What would be the advantage of merging? I like the idea of XMPP extension that don’t depend on a particular XMPP server.
Well, we built it as an extension that, in theory, doesn’t rely on any particular XMPP server. I was just wondering if we would have a better chance to get the community involved if we change the application to work as a plugin instead of a stand alone web application.
Cool. So it sounds like that we should go ahead and take the next steps then. Maybe upload the code to a public repository along with extensive, quality documentation? Also, it sounds like if we can alter the system to work with WorkGroup Queues and test using a couple XMPP servers we would have a higher chance of getting people interested.
Thanks for the advice! If you have any more, I would love to hear and to stay in contact. Otherwise, I guess, it’s time for me to polish off a few things and upload it.