Name: Duarte Barbosa
Contact/Email: duartebarbosa [AT] gmail [DOT] com, I'm pretty much always available there or on gtalk but I can take the dust of my ancient IRC knowledge if needed.
Background: I'm currently pursuing Masters majoring in Distributed Systems and minoring in Embedded ones. I've very nice skills with C and I'm pretty much used to read weird code. I speak fluent Java and I can script lots of stuff. I did some research with real-time linux kernel at INESC-ID and I've some experience with sysV-init files in order to reduce boot times although it may not be useful for this case. I'm nowhere the best person to work on this project but you can surely count on me and my passion.
Project title: SysV-init file creator from systemd service files
Project details: SysV-init is the de facto init system of any GNU/Linux system. It was conceived when multi-processing was future tech and it lacks many features which turn it pretty much obsolete. My objective is to conceive a tool which turns a bit easier the migration away from SysV init in a smooth way to systemd. Maintainers have a quite difficult time supplying both sysV-init scripts and systemd service files. This way, they would keep an (updated) systemd file and be able to supply in an automated way an "equivalent" sysV-init file which would be much less error-prone. My master thesis theme will be about system loaders so this is a nice beginning and obviously of my direct interest.
Synopsis: Tool to create SysV-init files from standard systemd service files.
Benefits to Debian: Right now, there are various distros shipping with systemd enabled by default. However, most of them don't provide a near as stable environment as Debian does. That's the reason it powers lots of old machines that can't crash. But we can't live on the past so this tool would provide a clean way to migrate more easily to systemd providing only a single file to maintain.
Deliverables: Application which creates SysV-init files
Project schedule: I believe this project will take some extra and unpredictable time due to constant systemd development. However, I can provide a simple schedule:
- 1. First 3 weeks - Design of the architecture of the application and study of both sysV and systemd.
- 2. One month - Implementation (I tend to always answer this with C, but maybe python is a better alternative)
- 3. Rest of the time - Debug. Although it may seem pretty straigh forward, I'm used to take more than the expected time on this phase. As this is supposed to have quality and be free of hacks I'll be devoting extra effort to this.
Exams and other commitments: I'll have exams on the first three weeks of June but with sparse dates. Although I expect to lower my productivity, I do think it won't be for much as the exams are pretty much far between.
Other summer plans: No... but I do plan to get out to see the sun (somedays).
Why Debian?: Since I started my academic life, I've been married with Debian. Started with a friendly derivative (Ubuntu) but soon I though that it was too bloated and not so developer-friendly as I had imagined and having previous experience (although short ones) with others distros, ok let's get Debian! Never looked back since then. I owe a lot to all the Debian team and the Free-Software. This is my (not the absolute first, but the absolute important first) contribution to the community which as welcomed me in such a lovely way.
Are you applying for other projects in SoC? At the moment, no. Will update this page if so.