zack's home page/ blog/ posts/ 2007/ 07/ uploaders vs maintainers

More Uploaders, Less Maintainers: Infrastructure Issues

When yet another discussion about "is collaborative maintenance good or not?" starts on debian- lists, the average lurker can easily identify two arguments / religions:

  1. it is good because duties are shared among co-maintainers

  2. it is bad because no one feels responsible for the co-maintained packages

Of the two religions, I stand for the first one. More generally I share the long term aim of decreasing as much as possible the "ownership" of Debian packages; collaborative maintenance is just a tool to achieve the goal.

In the meantime, I believe the second religion is supported also by technical deficiencies in some pieces of our infrastructure. In particular some of our developer tools artificially enforce a distinction among Maintainers and Uploaders where it wouldn't be necessarily needed. In the last few days I thought about some of them:

  • the BTS by-maintainer reporting mechanism. Years ago, I used to use the BTS as my personal Debian-related TODO list. When I had some time to kill, I just had to point my browser to http://bugs.debian.org/zack@debian.org and a lot of work I had to perform showed up. Today this is no longer the case: most packages of mine are co-maintained and only those with me as a Maintainer show up. Then I have to look for bugs belonging to debian-ocaml-maint@lists.d.o, pkg-vim-maintainers@lists.a.d.o, ... It's tiresome I'll never do that. Hence a request in #430986, shared by other: #225757, #310720.

  • dd-list is frequently used to report work needed by maintainers on -devel posts. Since long time it supports a '-u' flag for listing also maintainers, but it's used very seldom, probably just because people do not know about it. It's just a matter of changing a default and developers will find more easily their names in such lists.

Think about such deficiencies, I'm pretty sure we can spot and fix easily a lot of them in our tools.

Update: Lucas: I agree that the two arguments do not necessarily contradict, that's why I (initially) called them arguments. Then I moved to "religions", it was an oversimplification: there are of course pros and cons of collab maint. My main argument holds nevertheless: several perceived cons of collab maint are due to artificial distinctions maintainer/uploader in our tools, thanks for providing yet another example ... I want more! :)