Differences between revisions 35 and 36
Revision 35 as of 2013-03-14 07:16:06
Size: 8958
Editor: AndreasTille
Comment: Update some metadata of previous MoM projects; spacing
Revision 36 as of 2013-07-15 09:23:47
Size: 8958
Editor: AndreasTille
Comment: There was no sponsering in 04.2013
Deletions are marked like this. Additions are marked like this.
Line 42: Line 42:
||04.2013 ||AndreasTille ||MatyasASustik || inverse-covariance-estimate || to be decided || not yet || ||??.2013 ||AndreasTille ||MatyasASustik || inverse-covariance-estimate || to be decided || not yet ||

Mentoring of the Month (MoM)

In this effort the mentor dedicates a part of his spare time to a newcomer (the "student"), providing any packaging knowledge the Mentor has to enable the student to work more or less independently on packaging after passing this MoM period. The mentor will guide the student kindly into all secrets of Debian packaging with the example of a specific program, which is in the focus of the Debian Med team. The student is free to pick the package, however the mentor has a vetoing right in case he sees the package in some way "unfit for the MoM project" (too complicated, too non-free, too far away from Debian Med topic, too whatever). At the end of the month the goal is that

  1. The package in question is finished and uploaded.
  2. The student is able to do advanced packaging tasks and is introduced into the communication channels of Debian Med team.

The communication about this should be tagged with "[MoM]" in the subject line of the exchanged mails to enable others who might not be interested to procmail it out of their focus, and to enable other potential students to learn from this.

In this MoM process I would like to apply some strict rules:

  1. If the mentor is posting something on the mailing list containing [MoM] in the subject, the student tries hard to respond with a comment (if not better possible, something like: I have no idea but I will do some research like asking upstream or whatever - just leaving a hint that he feels responsible somehow).

  2. The student tries to follow any of the commits of the mentor to the package which is in focus of the MoM plan, and favorably sends a comment like
    • Its OK for me
    • What does this mean, please explain
    • I would prefer ... instead of your change

    To get informed about the changes the student needs to be subscribed to the commit mailing list or check the logs in the repository regularly (svn up | git pull; {svn|git} log). He or she should try to find a useful comment to any commit to make sure he has understood things and will be able to do it himself in the future.

  3. The student confirms that he succeeded in building the package according to the state in Vcs in case it builds, or that he is able to reproduce the error message of the build process.
  4. If there is a problem the student will ask on the Debian Med mailing list (tagging the subject [MoM] not to spam others too much.) This list should be the main communication channel to show others:

    1. we are working on a problem
    2. how things could be sorted out via open discussion
    3. how they could learn things about packaging

    Communication can also be done via the #debian-med IRC channel.

  5. The student posts a status report about the packaging every day, reporting about changes she did, issues she faced or discussions she did with upstream etc.

Please be careful. The MoM plan is work. While it is intended to show that work can be fun, the mentor expects the student to do some work in the same way as the mentor is offering a part of his spare time to do the other part of the work which needs to be done. So the mentor will ask the student whether he did his part of the work. If not the mentor will stop the work on this package and will continue with another MoM student and her package (in case there might be a waiting list).

The MoM effort was created in the hope that we will be able to train those silent observers of the list to become more vocal. This training not only provides technical knowledge. It turned out that this is only 50% of the job. The other part is communication which is heavily underestimated. By communication we do not mean plain chatting. It is about communication that leads straight to a technical implementation.

Current mentors, students and packages

Month

Mentor

Student

Package

Repository

Uploaded?

01+02.2012

AndreasTille

Luis Ibanez

fis-gtm

git://git.debian.org/collab-maint/fis-gtm.git

not yet (quite complex)

03.2012

AndreasTille

Scott Christley

libswarm

svn://svn.debian.org/debian-med/trunk/packages/swarm/trunk/

no

06.2012

AndreasTille

Eric Maeker

libquazip

svn://svn.debian.org/debian-med/trunk/packages/quazip/trunk/

yes

11.2012

AndreasTille

Tomás Di Domenico

python-csb

git://git.debian.org/debian-med/python-csb.git

yes

02.2013

AndreasTille

SukhbirS

hunspell-en-med

git://anonscm.debian.org/debian-med/hunspell-en-med.git

yes

??.2013

AndreasTille

MatyasASustik

inverse-covariance-estimate

to be decided

not yet

Students in the queue

  • Christoph Gille (Strap - svn://svn.debian.org/debian-med/trunk/packages/strap/trunk/)
  • Emmanuel Promayon (CamiTK - svn://svn.debian.org/debian-med/trunk/packages/camitk/ and related) -> February slot

For newcomers in the Debian Med team the most essential reading is the Debian Med policy document. Please put it under your pillow. It helps the student to in the initial phase when having first trouble with checking out via ssh as well as giving useful hints later.

Summary of past MoMs

15. January to 29. February 2012: Luis Ibanez packaging fis-gtm

Luis Ibanez took over the challenging task of starting with one of the very hard packages in the beginning. So the result that the package is not uploaded yet does not come unexpected. From a mentors perspective I'm happy with Luis because he learned basic packaging skills and communication channels inside Debian Med and did a pretty good overall job. He was quite verbose about his plans and success / problems so it was quite easy for me to guide him. For the package itself he started to rework the upstream build system using cmake which is probably a good idea anyway and helps other potential users outside Debian as well (which is a good thing). I will keep on helping to finalise the package once Luis will confirm success in getting cmake working.

1. March to 31. March 2012: Scott Christley packaging libswarm

It seems Scott Christley was not able to spend a comparable amount of time as Luis in the first MoM period. On the other hand Scott was not totally new to Debian packaging and there existed some preliminary packaging of libswarm in SVN. When he tried to finish the package it turned out that another precondition (libkawa) is needed and filed an ITP bug #666238. The plan is to finish this package first.

1. June to 30. June 2012: Eric Maeker packaging libquazip

In this MoM period it was the first time that the student finished the package ready for upload to NEW. To be honest the other packages were a bit more complex but this is quite a success for the MoM project anyway. Thanks to Eric for his effort in the MoM period.

12. November to 21. December 2012: Tomás Di Domenico packaging python-csb

Tomás also successfully finished his MoM period which was a bit more than one month but he did a really studious work checking several times also on python modules list to make sure his package will fit a high standard. Finally this is the case because he has choosen the hard way and fully support python2 and python3 in this package (as far as I know the first python module in our team approaching this). So he started as a beginner and ended up with an example package for other people trying to support both python versions. That's a nice job - thanks for your effort Tomás.

1. February to 28. February 2013: Sukhbir Singh packaging hunspell-en-med

Sukhbir has wisely choosen a package that is perfectly fit for a packaging beginner and managed to finish the package even after two weeks. The package is currently in the new queue waiting to be accepted by ftpmaster. Sukhbir has to be proven very careful because I even detected some problem with the spelling dictionary packages when beeing used in connection with OpenOffice.org/LibreOffice. The problem is not yet solved but he promised to work on this. A bug report should be filed once the package is accepted (it does not affect overall functionality and it seems it is even an "upstream problem". Thanks for the good work to Sukhbir - I'd be happy to see more contributions from yours to Debian Med.