Project ideas

Please read our philosophy for background information.


We intend that each student will be the project leader of his/her particular project and the only individual with commit privilege for the repository.

There will be one mentor and one backup mentor per project. In addition, the community will act as a rigid external reviewer of the student’s progress, as well as of the mentor’s.

General ideas

Our project ideas in general intend to focus on our innovative development environment from bare hardware to the graphical interface.

Concrete ideas

The listed ideas generally require good knowledge of C and experience with Unix-like operating systems. The difficulty ranges from medium to high. An academic background in computer science is desirable but not essential.

Suckless font rendering library

There is libdrw in suckless now, which still uses xft and fontconfig. Fontconfig and xft are ugly and require too much internal knowledge to be useful. The next logical layer in Linux evolved as pango and cairo. Both of course added HTML formatting and vector drawing. This is not needed to simply draw some text somewhere. And this is what a suckless font rendering library should do: Give it a font string and render at some position the given font without having to care about font specifics.

Requirements: C knowledge, some X11 knowledge and of course knowledge about the font formats and how to handle them.

Write ld wrapper or replacement for static linking

The GNU autotools such as automake and autoconf are completely unusable in non-chroot'ed cross-compile environments and often completely fail to produce statically linked libraries or executables. Also they are extremely slow and bloated.

The stali build system is not using autotools for good reason, however many Unix/Linux open source packages do. To create statically linked libraries out of the ld arguments we need an ld wrapper or re-implementation that creates static libraries or executables. This would enable us to build static libraries and executables out of any automake generated makefiles without the need to write make replacements or patching the build system of a particular package.

The ld wrapper needs to be extended to also link against uclibc first and if that fails to fallback to glibc, in order to produce smaller executables in the general case.

Requirements: Good C/Unix knowledge is essential, knowledge about linking/linker internals are desirable..

Write a decent bug and issue tracking system

We a need a decent ticket management system, as this is a common task in today’s daily activity in business and private time management. This task includes the design and implementation of a suckless system that solves the problems of TTS (Trouble Ticket System), ARS (Action Request System) and IRS (Incident Response System) all together. It also must be usable as a bug tracking system.

For now suckless is using a mailinglist which fits all needs. Beware when you find the new and innovative way to report bugs that this is reality.

Requirements: Good C/Shell/web technology/HTML knowledge would be desirable, knowledge of bug tracking and issue tracking in practice is essential.

Write a decent mailing list Web archive system

All web archive systems such as hypermail, pipermail, etc. have plenty drawbacks and are quite out-dated. This task requires to write a completely new web mailing list archiving tool that follows the thread view concepts found in the mutt MUA and which is designed with low footprint and efficiency in mind.

We expect this tool as a stand-alone Unix tool written in C or shell. To get started you could use Dovecot to produce a sanitized structure:

printf "1 select inbox\n2 thread references us-ascii all\n3 fetch 1:*
envelope\n4 logout\n" |
/usr/local/libexec/dovecot/imap  2>/dev/null

Requirements: Good C/Shell/HTML knowledge would be desirable.

Improve sltar

sltar is a simplified tar implementation which lacks gzip and bzip2 integration. The task requires extending sltar with support for these compression applications and writing a test suite for it.

Requirements: Good C knowledge would be desirable.

Write cookie handler for surf

The biggest disadvantage of surf is sloppy cookie handling. libwebkit and libsoup (which are used for HTTP) were never designed to run in multiple processes simultaneously.

This task requires writing a new cookie handler in surf which:

Requirements: Good knowledge of C and POSIX file locking. Basic knowledge of GTK and its other evil friends.

Gopher services

Gopher is a sane protocol which has hierarchy in its design. It allows the abstraction of a mass of information in a filesystem. The goal of this meta project is to find ideas how to implement gopher services to easily access the web and new information.

See the protocol for how easy it is to write a menu, which can be seen as a directory.

Anyone creating a gopher interface to will get a bonus.

Requirements: Just some shell scripting and a way to setup a gopher daemon is required. Everyone can do this.

A sane backend for surf

There is dillo, netsurf and abaco which implement HTML. The problem is Javascript and extension to replace webkit as the big dependency hell for web rendering in surf.

If you prepare to work on this project, plan ahead in recruting more developers. You will need them.

Requirements: Very good C knowledge, a very good knowledge in web standards and how to strip them down to the suckless level.