Re: internship on the Hurd

[Jeff Bailey]

On Tue, Sep 17, 2002 at 06:21:06PM +0200, Roel Neefs wrote:

> > While it's obviously important to work on something you're
> > interested in, we have a whole host of things that could use work
> > that keep us from using the Hurd everyday -- Especially if you
> > have a one year internship to dedicate to it!

> The intersnhip itself is only 3 months, afterwards i have to write a
> thesis. But if i like it i will probably keep working on it. :)

> > It would help if you could post some notes about your experience
> > and background, so we don't give you suggestions that are either
> > too easy to not be worth it, or too hard for your skills and
> > experience.

> Well, i have 4 years of experience in scheme. :) 3 years
> C/C++/java. I have never worked on kernels, but i'm a fast
> learner. (reading a lot of books about it at the moment) But what i
> would work on would need to be worth an internship, i don't think
> porting programs to the hurd will be allowed by my professor.

The 'TODO' file in the root of the Hurd source is a very good place to
start, but there's also some other things that might be doable as
three month projects:

o Write test suites for the Hurd, or the hurd and mach subtrees of
glibc.

o Hack inetutils to run all of its daemons with no priviledges on
startup.  (If that's too simple, go on to figure out how to make it
work with Kerberos and other password-less authentication schemes).

These both involve developing (and possible fixing) parts of the Hurd
that make us clever and special.  They also both can be either short
or long projects, depending on how thorough you want to be.  They also
have very short returns so you don't have to worry about something
going wrong and having nothing to show for 3 months of work.

Tks,
Jeff Bailey


-- 
learning from failures is nice in theory...
but in practice, it sucks :)
 - Wolfgang Jaehrling

 LocalWords:  Jaehrling