[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [RFC] Measuring skills of a Debian Developer

On Mon, Nov 13, 2000 at 11:14:31PM +0100, Martin Bialasinski wrote:
> Debian is a volunteer efford. I do the things I like and am interested
> in. I need noone assigning me things to do. And there is no need for
> me to "register" to be allowed to work on some part of the project.

That's the idea. But you need to register your ITP's before you package
something right? This is just an innocent extension. I will explain the
idea below... Please read on.

> I study economics. One specialisation I do is organisation theory. I
> recognise your intention, and all I can tell you is that it doesn't
> fit to Debian at all.

IMHO this would be more related to sociology than undergrad organisation
theory. It is also a field where no seminal work is present.
I'd urge you to go ahead and find some papers on peer-review systems
if you'd like. Today, I think there's no really great peer-review
system that works on interconnected networks.

Let me tell you about an idea now, and you can then decide whether it
fits to Debian.

The idea is making things in the open. Why is that? Because, as you will
have noticed we are interested in the work done rather than authorship!

If people have a way to electronically examine how projects are going,
then they can use their volunteer time more efficiently, and more people
can work on the projects they want. That's one aspect. Another aspect
is about automation. Now, of course you can talk to people and then
work on the projects by, you know, using this mailing list. But the
point is that mailing lists aren't an efficient and reliable way to do that.
That's why there's BTS and WNPP in the first place, you can point out to
better examples I'm sure.

How do we do that? What kind of an infrastructure is suitable for that?
Ultimately, I'm convinced that this is something that we wouldn't be
able to figure out in a short time. It's probably a difficult problem.
But of course we can make a minimal set of infrastructure tools to aid
in development.

The thing is that the system should not care who it is dealing with.
So basically, there won't be "someone" assigning you "a task". You'd
just register if you're looking for some project to work in your volunteer
time. As I said, like the GNU task list. You shouldn't have to register
or it shouldn't be a strict requirement to get things done. [***]

A nice set of tools could be:
  * modify wiki wiki web (is that right?) so that it supports
    task-volunteer matching. Basically, every task has its pages,
    and developers can modify those pages if they're logged on. Everything
    is publicly viewable of course.
  * private usenet groups to aid in task specific conversations, the
    discussions are archived, and a web mirror to the page is provided.
  * task-volunteer thingie itself. Developers login, and they have a
    profile that they can edit. They basically talk about themselves
    what they've done and what skills they've got. The thing is that
    there're a these skill categories that are related to free software
    and in somewhat more detail than like
      Which skills do you have? Check all that apply.
      o C/C++
      o Java
      o Perl
      o Kernel development
    You know, this is simply not good enough. Assume that people are
    looking for help on an autoconf/automake build system for a project.
    Now, it would be nice if they could say, "we need someone...". It'd
    be even better if they could search for one.
  * now about the task-volunteer matching. The system just looks at
   what tasks are to be done, and what skills they require and then assign
   people (who are not busy) to those tasks. It tries to do load balancing,
   so it tries to give people equal number of tasks...
You know what I'm thinking of is like a "free" version of cosource/
sourcexchange stuff... With more juice of course. What I talk about in
this mail is extremely simple minded, I'm aware. But the idea can be
taken and put into good use.


[***] Very important: there should be no human administrators/managers
who decide about people. That would be contrary to the spirit of Internet.
Of course there would be people maintaining the software, but that's
a different thing.

Eray (exa) Ozkural
Comp. Sci. Dept., Bilkent University, Ankara
e-mail: erayo@cs.bilkent.edu.tr
www: http://www.cs.bilkent.edu.tr/~erayo

Reply to: