Desktop Integration of SIP/IM in the (K)Ubuntu Desktop



Existing free IM software connects to either a single network (such as Jabber) and implements as much of that network's functionality as possible, or provides plugin architecture to connect to various networks and attempts to implement as many of the features in common on those networks. In either case the result is a large base of code which tries to do a great number of different tasks from the same program, leading to complicated and confusing user interfaces, poorly maintained areas of code, and violates the general principle of 'do one thing and do it well'. Furthermore, each monolithic program tends to export its own unique IPC interface for interacting with other programs, meaning that if integrating its functionality into other applications is desired, the work must be done for a specific application, restricting interoperability and the user's choice.


Todays popular desktop applications are e-mail clients, IM clients and VoIP phone software, and they each open their own connections to their own servers, and provide their own different UI to similar functionality like having a text conversation, having a voice call or transferring a file. The existing Galago project set out to define a library, server and D-Bus protocol to share presence and address book information between any application that deals with people, and is already gaining support in existing IM and e-mail client software (eg it provides plugins that allow you to see in Evolution or Beagle that your contact is online to talk to in Gaim). Extending this idea by defining an extensible interface between backend IM & VOIP server connections and the frontend UI programs will allow seperate simplified programs to implement one part of the overall IM & VOIP functionality, and increase the code shared between different client UIs. The standardised interface will allow the shared backend connections to be available to the best application for the task in hand - a phone application for making phone calls, an IM application for text chat, and a file manager for sending files.

Scope and Use Cases

1. Ken installs Ubuntu and wants to be able to write an e-mail, chat with instant messages or his friend Barbie. He starts up his address book and adds Barbie's e-mail address, IM username, SIP ID or phone number (or maybe just a single Launchpad ID to do it all? :D), and it tells him that she is on-line and available for an IM chat or to recieve a VOIP call. Ken clicks on her name and his IM client starts for him to talk to her.

2. Barbie is talking to her friend Kelly over IM and the buddy list shows that she is also available for a voice call. She clicks the icon and the phone application starts and sets up a call with Kelly.

3. Ken is browsing his photographs in his file manager and finds a photograph of himself and Barbie on holiday in Spain. He right clicks and selects "Send To" and gets a list of all his contacts who are on-line and can recieve files. He chooses Kelly and the file transfer is initiated.

Implementation Plan

We will work closely together with the Telepathy project and use their D-Bus protocol for communicating between the back-end connection managers and client front-ends. The overview (slightly outdated) shows how different UI applications can share the same connection managers to achieve their respective tasks, and query a presence manager which is responsible for aggregating the availability and capabilities of your contacts via the various connections.

Prototyping of a Jabber connection manager and a text mode client is currently being done in Python by the Telepathy team and they will publish their protocol specification shortly, and by the end of the year they plan to have implemented native SIP and Jabber connection managers, a client library, and a feed of Telepathy presence information into Galago. After the specification is published, seperate connection managers can be implemented to support other proprietary protocols such as MSN, Yahoo, AOL and ICQ, using the best-of-breed free software libraries available for connecting to these services (many of these are written in Python, so this work can be based off the prototype Jabber connection manager).

IM and VOIP UIs can be written or existing applications modified to make use of the Python code or the native Telepathy client library to make connections using the connection managers, carry out IM sessions and make and recieve VOIP calls (the Telepathy project will use the Farsight library for this). Nautilus/Konqueror/gthumb/F-Spot can be patched to easily initiate file transfers with people in your address book.

The Galago presence manager has already gained some acceptance for sharing presence information between applications in the GNOME desktop and provides a high-level API for dealing with people and presence information. It can easily be integrated with its equivalent KIMProxy functionality in KDE, or a seperate Telepathy feed can export the same information over KIMProxy.

Cooperation with other projects

Gnomemeeting's dbus interface is currently being reworked... before that, the new interface has been presented for comments on gnomemeeting-devel-list and the dbus list on Gnomemeeting has audio, video, H.323 and SIP ; it is a mature VoIP client. How come no serious contact has been made yet ? [written by Snark the 20051107]

[CategoryUbz] [CategoryMOTUIM] [CategoryMOTU]

MOTU/Teams/IM/DesktopIntegrationSIPIM (last edited 2008-08-06 16:41:24 by localhost)