Linux Packaging project

From Kiwix
Revision as of 06:19, 28 August 2012 by Kelson (talk | contribs) (→‎People)
Jump to navigation Jump to search

The Linux Packaging project is a project to improve GNU/Linux integration of Kiwix.

Properties

  • Legal name of organization or individual requesting this grant: Emmanuel Engelhart
  • Project contact name: Emmanuel Engelhart
  • Project contact username or email: Kelson
  • Project contact title (position): Project manager
  • Project leader name: Emmanuel Engelhart
  • Project leader username or email: Kelson
  • Project leader title (position), if any: Project manager
  • Full project name: Kiwix Linux Package Project
  • Provisional target start date: 01-01-2013
  • Provisional completion date: 01-06-2013

Goals

Kiwix is a software which works and although we still try to improve it, its big challenge is currently to be known by people. That why we need to bring Kiwix to people and organization as near as possible.

On Windows, most of the work we need to do is communication work, people are used to donwload software directly from the Web, so we still need to convince them to do it. But, with Linux distributions, this works differently: it's difficult to install softwares which are not in the internal software catalog (or called "official repository"). That's why is really important to have Kiwix in those software catalogs for the most used Linux distributions, and this is the main purpose of this project.

Here are the details:

  • Package for Ubuntu + official repository
  • Package for Fedora + official repository
  • Package for OpenSuse + official repository
  • Package for Archlinux + official repository
  • Package for FreeBSD + official repository
  • Packages for openWRT and DDWRT

Activities

This project would be split in 3 work-packages:

1 - Compilation (3000 CHF)

  • Integrate Mozilla source code to Kiwix source code
  • Compile Kiwix against Mozilla source code (avoid using Xulrunner and Xulrunner SDK)
  • Check that everything compile and run fine like that for what we already release and support (Sugar, WIN, OSX, static&dynaimc + debian package)
  • Integrate the work in the compilation farm
  • Everything should work as good like before

2 - Package creation (2500 CHF)

  • Create the packages for the recent version of the listed GNU/Linux distributions
  • Create the packages with the compile farm
  • Check that everything works fine

3 - Package integration in official distribution repositories (2500 CHF)

  • Take contact with packagers, find an official packager or do it yourself
  • Do the necessary to integrate Kiwix to the distribution, at least in the testing version (experimental/testing for Debian)
  • Post-project, do the necessary that everything run fine and that Kiwix finaly land to the current release repository

4 - [Open/DD]-WRT Integration (2000 CHF)

  • Create the packages (one click install using the router WEB UI) for those two WRT distributions
  • Muss work with the most common model using this distributions (mips architecture?)
  • Package must have an easy interface to chose/change the content to spread, configure the interface (local network/internet)) and other mandatory things
  • Take care that this package will be integrated in official catalogs of thus distributions

Resources

  • 680 hours of professional development
  • 80 hours of volunteer time for the project management
  • 5 months of time
  • 10.000 CHF

Fit strategy and user demand

Increase reach is currently the first Wikimedia strategic goal. Offline is an essential part of the Wikimedia strategy to increase the audience of our content, and Kiwix as official Wikimedia offline reader is the offline lead tool.

GNU/Linux represents only a small part of the audience, but this is the audience we want to push. We think that free software is the best solution to access to free content. This is also the best solution for economically poor people who are our most targeted audience. At the same time, this part of the audience is used to collaborate, and so one often make more feedbacks about the software - something essential to improve it.

Currently, around 20.000 people per month download Kiwix. This is a lot and this number is growing constantly. But we still things that millions of people need Kiwix and that in fact this number represents only a small fraction of our potential audience. That's why such project are essential to push the software to those users.

Wins

  • Fully deployment of Kiwix in the most known and used GNU/Linux distributions
  • Remove the Xulrunner dependency which is an almost unmaintained piece of software
  • Improve the global quality of Kiwix compilation scripts
  • Increase audience and quality of feedbacks
  • Create the packages to use kiwix-serve for the time with routers (widely used and really cheap). This would be an improvement of the Afripedia project.

People

This project would be led by Kelson and executed by the company Yeleman which has a long experience with Kiwix codebase.

See also