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

Re: Proposal: Recall the Project Leader

Clint Adams <schizo@debian.org> writes:

> Now, if you become the release manager, and your employer makes your
> compensation contingent on Debian not releasing before February of 2010,
> no one can NMU the release.  Theoretically, we could replace you, but we
> cannot fix the problem directly.

> Would you not agree that this affects the risk assessment?

If I became the release manager and some other distribution offered me
$50,000 if Debian doesn't release before February of 2010, the situation
is the same.  What you're talking about here, in my opinion, is a simple
question of ethics.  The conflict could come from any number of sources,
including ones that aren't even monetary, and the risk is present whether
I'm being paid to work on Debian or not since it doesn't have to come from
my employer.

The solution to this sort of situation is, again, a matter of ethics.  As
a Debian Developer, I agreed to be part of this project.  To me, that
carries an ethical obligation to make decisions for the general good of
the project.  Should I be put into a situation where I don't feel like I
can do that without conflicts of interest, I would recuse myself.  Should
someone offer me a bargain like the above, I would refuse it.  Should the
Debian project as a whole not trust me to act ethically, it shouldn't
trust me with that sort of position.

I don't think that this is so dire of a peril that we should throw up our
hands, decide no one can be trusted to behave ethically in such a
situation, and bar any employer support of any position we think is
important (and that line is hard to draw and isn't just delegates, since
the real risk is more subtle than simple vandalism of the type that you
described and the gcc maintainers, kernel maintainers, d-i developers,
libc maintainers, and so forth all also have important, central roles in
Debian).  These sorts of ethical requirements are simply not that
uncommon, and the vast majority of people negotiate them without any major

The best thing the project can do to help with this is to work to avoid
small points of failure and to put more people in a position to help
should a conflict arise.  We *have* had a problem with this, and I *do*
think it's a problem.  Ironically, release management is one of the areas
where this is *less* of a problem....

> If my employer encourages me to spend an hour a week working on Debian,
> I think that's fine.  If my employer demands that I spend one hour per
> week trying to get HotJava through NEW, I will either refuse or resign
> myself to be an unscrupulous hypocrite.

Exactly.  And I expect any ftp-master would do the same thing, and that
anyone who wouldn't make that decision shouldn't become an ftp-master.
And as long as they held that ethical position, it wouldn't matter whether
their employer encouraged them to spend an hour a week working on the NEW
queue (thus making them funded to work on Debian).

>> Furthermore, to address this from a more personal level, I *will not*
>> segregate my life to that degree.  I find it deeply unpleasant to do so
>> and have no particular desire to be involved in a project that would
>> require me to make sharp distinctions between work time and not-work
>> time.  I carefully chose my employer precisely so that I would *not*
>> have to do that.  I chose an employer that would not create difficult
>> conflicts of interest, I have structured my life so that I can work on
>> what I feel needs to get done without having to worry about who "owns"
>> that work, and I would deeply resent someone trying to interfere with
>> that *personal* decision.

> Then I'm not sure of what I might be saying that applies to you.

Basically, I'm trying to make the argument that other people are just as
capable of doing this as I am, and that it's possible to negotiate these
waters without creating conflicts of interest.  And that doing so is very
common in projects of this kind.

>> I don't consider that particularly relevant.  The developer body is
>> large enough that it's unlikely to be united in anything.

> Perhaps this is nostalgic hagiography, but we used to be united in
> producing a quality OS.

*heh*.  Touché, and a fair point.  I probably should have instead said
that the developer project is large enough that it's unlikely to be united
in anything other than the sorts of general goals without which they
wouldn't bother to be a Debian Developer at all.

> I suspect that I've reached Matthew Wilcox's 3-post-per-day limit now.

I managed to teach lintian's dependency handling how to understand that
a|b implies a|b|c, so I'm giving myself a (small) allowance of time to
respond to d-v.  But now I need to go fix serious bugs in gnubg.

Russ Allbery (rra@debian.org)               <http://www.eyrie.org/~eagle/>

Reply to: