This would enable any website owner to add instant messaging functionality to their website, and due to the federated nature of XMPP, users could chat to any other public XMPP account (once they have been accepted as a contact).
One thing that previously prevented you from using converse.js on its own, was the fact that it made XHR calls to the Plone backend to fetch user data. To fix this, I added vCard support, to converse.js but also to Plone by adding it to collective.xmpp.core.
Last week I reached a significant milestone on the path to this goal, and I'd like to take a moment and share with you.
It is now possible to use converse.js (in a static HTML page) to communicate with Jabber accounts on any public server.
In the demo below, I illustrate this by chatting to a Google user and to a Jabber.org user. In the converse.js page, I'm authenticated with a Jappix.com account and I also use their connection manager to connect to their XMPP server. If you're doing XMPP via HTTP (i.e in the browser), you'll need a connection manager as a bridge to your intended XMPP server. Thanks a lot to the Jappix guys for making their connection manager public!
Note: For detailed fullscreen, make sure that video quality is set to HD, by clicking the gear icon on the video player.
UPDATE: Since this blog post, a production ready version of converse.js has been released. So the following is no longer applicable.
I'm sorry to say that converse.js is not yet 100% ready for primetime as a standalone JS app. There are a few more hurdles (and nice-to-haves) to overcome before we'll achieve this goal:
- Searching and adding new users still does XHR calls to Plone and should instead query the XMPP server.
- Service discovery support (e.g so that multi-user chat is only available for servers that support it).
- Some kind of "Keep me logged in" support when users log in manually.
- I'm not proud of the CSS and it could probably be improved upon quite a bit.
- The Jasmine tests are out of date and not passing, also more tests are required.
If you are interested in the project, please contribute by forking the code on github.