Summer of Code 2006 Project Ideas
Here are the project ideas that were suggested for the Summer of Code 2006 ...
- 1 Client
- 1.1 Coccinella: an aquarium for kids
- 1.2 Psi
- 1.3 General Client Enhancements
- 1.4 Java Jingle
- 2 Server
- 3 Federation
- 4 Transports / Gateways
- 5 Development Tools
- 6 Jabber Enhancement Protocols
- 7 Non-IM Applications
- 8 Jabber Related Projects of Other Mentoring Organizations
Coccinella: an aquarium for kids
- Mentor: Mats Bengtsson
- Programming language: Tcl/tk
- Coolness factor: high
This project involves the creation of an interface for kids to Coccinella. You need to create a new roster style, the fishy roster style, with an aquarium containing fishes that represents the contacts (like ralphm's Jabber Aquarium). The fishes need to be be animated with User Mood. For example, when a contact is happy, the fish will swim quickly in the aquarium with a smile on his face. The graphics will be done in 2D, but when time allows, they can be made in 3D.
- Creation of educative whiteboard games suitable for kids.
- User Activities (see next section).
- Message mark-up (XMPP-IM).
- Creation of flashy emoticons like MSN.
This project will involve mostly new code, and less boring code rewriting.
XEPs you will/can work on
Psi was the first Jabber client to release a branch with support for voice calling using the libjingle library released by Google. The main Psi branch has since migrated to the Qt4 toolkit (the Jingle branch uses Qt3) and libjingle has moved on several versions. A student could merge the psi-jingle branch changes back, and a newer libjingle, into the main branch, and finish work on providing media interfacing on non-linux platforms. Edit: The Psi developers have since ported and merged the Jingle support into mainline, provided OSX support and are working on Windows support, so an interested student would need to find ways to extend this.
Link-local messaging allows clients to communicate directly, without the use of a server. The libraries used by Psi support the LAN discovery part of this, but not the chat protocol which lies on top. An interested student could implement the chat protocol in the library, and then integrate this in the Psi GUI.
There already, through Jingle-audio, exists a protocol for voice calling over xmpp, which Psi already supports (in a seperate branch, see above); a particularly adventurous student might wish to extend this to video support for video calling. A student interested in this project would have to interface with video hardware and sound hardware (probably re-using the libjingle sound code), and provide streaming capabilities, probably through libjingle. A negotiation protocol would be developed, although assistance would be given with this, and it is not the focus of the project.
Several approaches to whiteboarding have recently been proposed for Jabber, revolving around the SVG standard. A student interested in this could prototype a method for encapsulating changing SVG data in the xmpp stream and implement an editor/viewer in Psi. Also linked to the "Sharing Animations" task.
Linked to whiteboarding, but a less gargantuan undertaking, would be implementing a method for sharing progressive SVG animations between Psi instances, and rendering them within Psi. A student interested in whiteboarding would be advised to start here and, indeed, this is a worthwhile project in its own right.
Interesting Chat Widgets
The chat view in Psi is currently fairly plain, while other clients have far 'whizzier' interfaces. This is an open-ended project to attempt to 'make the chat widget more interesting', starting with finishing xhtml-im support. There are many directions this project could go, one of which, although possibly not the most interesting, would be to support the same theme files used by Adium and Google Talk.
Psi has long been in need of a plugin interface. A framework for loading plugins has already been worked on, but the interface itself between Psi and plugins needs to be defined. Such an interface must completely abstract itself away from the internal workings of Psi and Iris, allowing a stable plugin system, even when Psi's code changes. A student could finish the plugin management and implement an interface, bringing plugins to Psi.
A Mail-Like History
Psi's history behaviour is somewhat basic at present. An interested student could overhaul this interface, providing something more akin to a mail client, with ordering, sorting and searching based upon contact, time, content etc.
Psi has in place some unit tests to ensure that changes to the codebase do not introduce errors to the code. A student could extend these tests across the rest of the codebase.
Any Other Business
There are many other possible projects involving Psi. Any students with their own ideas of things that can be added or improved can discuss their ideas with the developers before applying if they wish.
General Client Enhancements
Pick an existing client, and implement a full XEP or improve what's present.(proposed by User:Zion)
Some places to start:
Few clients support the whole thing.
It's experimental, but we currently have a chicken and egg problem with PubSub in general; we need clients to implement it to encourage servers to expose it. Experimental implementations include those in Psi and a patch for ejabberd.
It's experimental too, but implementing experimental XEPs might speed up proceedings of those XEPs to get draft or final state.
Create a Jingle implementation in Java and JavaSound. Jingle is an XMPP extension that primarily (for now) enables VOIP. A Java implementation would spread the protocol more widely. It's also very useful to have several implementations for testing purposes. (proposed by User:MTucker)
- Add migration scripts from other IM servers to Wildfire. This could include non-XMPP and XMPP servers. (proposed by User:MTucker)
- Create a SIP/SIMPLE gateway to make it easier for users to fully transition to XMPP. (proposed by User:MTucker)
- Perform integration with other telephony systems, like Asterisk-IM does for Asterisk integration. (proposed by User:MTucker)
- Contribute to the Pampero project, which will bring massive scalability to Wildfire. (proposed by User:MTucker)
- Implement XEP-0079: Advanced Message Processing and XEP-0133: Service Administration. The server already supports Ad-hoc commands and a few services defined in XEP-0133 but there is still a lot of work to do. (proposed by User:Gaston)
- Implement XEP-0163: Personal Eventing Protocol. Wildfire supports latest version of Pubsub that is a prerequisite of PEP. Moreover, the original design of Pubsub considered that different pubsub profiles could be created. However, there are still important changes that need to be done to support PEP. (proposed by User:Gaston)
- It would be great if the ejabberd server (written in Erlang) supported the latest version of Publish-Subscribe including the Personal Eventing subset. (proposed by User:Stpeter)
- Support for the HTTP Binding would be great but I think the folks at mabber have written this and just need to contribute their code. (proposed by User:Stpeter)
There already is a draft implementation, is it still worth to list this here? See bug 25.
- Create a Jabber to Jabber transport as a module for ejabberd. Jabberlang will be useful.
- Create a system to allow people to migrate easily from Microsoft Office Live Communications Server (LCS) to ejabberd. LCS lacks many important features amongst XMPP support and very high scalability which makes this migration tool very useful. It would be great if migration from LCS to ejabberd is so easy and perfect that it outperforms the installation process of LCS. In other words: migrating to ejabberd should be more easy than installing LCS.
- ejabberd already supports client connections via XMPP and via HTTP, it would be nice if IRC is also supported. So, create a connection module that:
- Allows users to connect to ejabberd with their favorite IRC client, just like ejabberd is a normal IRC server.
- Allows users to enter Multi-User Chat rooms and enter IRC channels on another server via the IRC transport that ejabberd already has.
- By default restricts access to users that connect via IRC in a similar way as SASL Anonymous.
- Allows users to register their IRC nickname, this will automatically create a Jabber ID and will eventually enable a less restrictive environment.
- Allows users to change MUC room topic, retrieve information about users (vCard, client version,...), and much more.
Jabberd14 is a commonly used Jabber server implementation that servers many sites. Thus support of the latest XEPs would be a great feature to it. Examples are:
- the latest version of Publish-Subscribe (XEP-0060)
- Personal Eventing Protocol (XEP-0163)
For XEP-0060 either a current implementation (e. g. Idavoll2) could be updated or a new implementation be written, depending on what's the best choice. (proposed by User:Maxi)
Generic Server Work
- A lot of servers claim XMPP-compliance, thus following the RFCs strictly. A test suite for interoperability between clients and servers can ratify these claims. (proposed by User:Astro) Test suite coverage over major XMPP extensions for protocols like MUC, pubsub, etc would be highly useful.
- Contribute to an existing XMPP load testing project or create a new one. A test suite that can easily work across many client machines would be ideal for testing very high load. Suggestion: Java implementation so that it's cross-platform. (proposed by User:MTucker)
Why re-invent the wheel? There already is Tsung:
- multi-platform and based on an open-source language
- Jabber benchmarking (amongist others)
- clustering features of Erlang
So, I guess it is better to add features to Tsung. I think they have a TODO list... (comment by User:Sander)
Distributed monitoring of the Jabber network
(proposed by Lucas Nussbaum - xmpp:email@example.com)
The Jabber network isn't really known for its reliability, and administrators don't always have the necessary tools to monitor their server correctly. The data provided by http://public.jabbernet.dk/mrtg/ is not enough. The goal of this project is to build a testing framework to be able to monitor the public servers and answer those questions :
- do client connections using SASL, TLS or SSL currently work on server S1 ? (not only checking if the TCP port is open, but full login)
- do server-to-server works between server S1 and S2 (by exchanging ping messages on XMPP) ?
This would allow to build a more detailed status of the Jabber network. Additionally, it would be nice to be able to add a contact email, so the administrator can get informed of his server's problems.
(addendum by User:Stpeter)
Integrating this into the XMPP Federation database and website would be super.
Transports / Gateways
Jabber Enhancement Protocols
Jabber authentication to CMS and blogs
(proposed by User:Halr9000)
See my XMPP and CMS post on the topic.
(proposed by Wikipedia:User:Henna)
Jabber Remote Control of VLC Media Player
Write a plugin modul for VLC Media Player which simply acts like the HTTP remote control interface. Use XEP-0050 for providing functions like play, next, previous, etc. so you are able to control VLC Media Player via Jabber.
(proposed by User:Tobiasfar)
Jabber Related Projects of Other Mentoring Organizations
- Adium X "Improve Jabber support by integrating (and possibly improving) a Jabber library. Possibilities include telepathy, libjingle, Smack, and improving LibGaim."
- Handhelds.org "IM Client for Opie/GPE integrated with phone and PIM"
- Haskell.org "XMPP (aka Jabber) bindings for Haskell"
- KDE "Improve Jingle video/audio support and Audio/Video device configuration", "Jabber full MUC support.", and "Whiteboard support for jabber."
- LispNYC "Implement a Google Talk (Jabber) chat client for McCLIM."
- Open Source Applications Foundation "Jabber in Chandler"
- The Free Earth Foundation "Student-Teacher Interaction System"