Re: [gentoo-dev] Google SOC 2008

[Alec Warner]

On 2/26/08, joshua jackson <tsunam@g.o> wrote:
> Rémi Cardona wrote:
>  > joshua jackson a écrit :
>  >> 2) We need mentors, so far confirmed I have: Diego and Saleem
>  >
>  > What kind of work is involved there? I wouldn't mind being a mentor
>  > but I'd like to know a bit more about what's expected from a "good"
>  > mentor.

I hope to have documentation up by the end of the week.

>  >
>  > Thanks,
>  >
>  > Rémi
>
> There's a few requirements. Being decent in a language or multiple
>  languages would certainly be a plus, as the students are writing code.
>  having someone writing something in C or C++ when you've never touched
>  it wouldn't exactly work out that well obviously.

You will reviewing the student's ideas and work.  You will need
whatever skills are required to to that; it will depend highy on what
project you choose to mentor.

>
>  Having time to interact with the student as well. They are getting paid
>  as are we as an organization, so helping them and giving them idea's is
>  needed. Touching base and making sure they are still progressing on
>  their projects. Its summer...but they are being paid to work so it is a
>  job as I hope as a mentor you would take a similar approach as well.

You should expect to spend a minimum of five (5) hours a week doing
reviews of your student's code.  Ideally you would make remarks for
readability and documentation as well as design and implementation.
You should meet with your student regularly to discuss progress; if
the student is having problems they should be brought up and dealt
with.  You should run your students code to make sure it functions as
he or she said it would.  If it has tests, run those too.  If it
doesn't have tests; ask the student to write some ;)

>
>  Basically you act as a technical boss/mentor/leader to someone.

I would argue that you are a peer, not a boss.  In the end, someone
the student should probably take advice from in order to have a
decently working product by summer's end.

>
>  One area I'm working on for this year is goals within the overall goal
>  that can be implemented hopefully even if the project isn't fully
>  finished before the end of the project.
>
>  For the mentor it won't be a full time commitment but having time to
>  talk with/help someone with their project and explain how to implement
>  this feature or get this information from Gentoo's system would take
>  some time.
>
>  As a warning some people come on board with excellent programming
>  ability that you don't need to do much with/for. Some well its their
>  first real jump into programming and need more attention. I plan on
>  having a few additional questions for the Gentoo related applications to
>  help define that kind of thing and hope to get people together in such a
>  way.

While true I don't think it is fair for the 'superstar' guy to just
let him do whatever.
*review their code*.  Read it, run it, see what it does.  If the
student says they wrote it in 3 hours and it is awesome and well
commented try running it through a profiler; find places where it
could be improved.  Look for poorly implemented algorithms, memory
leaks, over-use of memory etc...

>
>  As well a few of the people who are helping run it have agreed to help
>  poke/get updates from the mentors and step in and help in whatever way
>  possible to the students as well. Basically, we want as many projects to
>  succeed as possible and become a daily tool or vastly improved tools as
>  possible. Its certainly something that is possible.

My ideas are all insane (I probably only did half when I was a mentor)
but I want the students to have a good experience and I want the code
to not suck and that requires a mentor who actually gives a damn and
is willing to put time into it.

-Alec

>
>
>  --
>  gentoo-dev@l.g.o mailing list
>
>
-- 
gentoo-dev@l.g.o mailing list