LWN: Comments on "Possible changes to Debian's decision-making processes"

Possible changes to Debian's decision-making processes

marcH — Mon, 25 Oct 2021 05:29:08 +0000

> The name Debian brings to mind a Linux distribution, but the Debian project is far more than that; it is an ongoing experiment in democratic project governance. [...] Debian's voting scheme allows a ballot to contain numerous options with small differences without fear of splitting the vote in a way that causes a relatively unpopular option to ultimately prevail. At its best, it creates ballots where developers can vote for the options they want rather than just voting against the worst case.

Meanwhile, in the real world:
https://www.nytimes.com/2021/09/23/podcasts/the-daily/red...
> Clearly aware of the stakes, New York Democrats are considering a tactic that is usually a preserve of the Republican Party: gerrymandering.

Possible changes to Debian's decision-making processes

flussence — Fri, 22 Oct 2021 19:40:06 +0000

That was a great explanation, thanks! Much clearer to me now.

Possible changes to Debian's decision-making processes

NYKevin — Wed, 20 Oct 2021 18:43:24 +0000

You can essentially think of their voting system as a topological sort over the graph of preferences. If a majority prefers X over Y, then draw an edge from X to Y. Then, hopefully, you have a DAG, and if you do have a DAG, you should* have just one source vertex, so hopefully, you can just pick that one source vertex as the winner. This is not very complicated all by itself and can be simply explained as "pick the option that would beat everything else in individual runoff elections."

Most of the complexity comes down to "what if there's a cycle?", in which case you have to drop one or more edges until there isn't a cycle anymore. Debian's system is designed to drop edges in such a way as to invalidate the lowest total edge weight overall (where the weight of an edge is the total number of voters who prefer one option over the other). There are numerous other ways to do it, and you can go off into long tangents about the mathematics of it all, but suffice it to say that 1) there's no "perfect" option (for a rather technical definition of "perfect") and 2) Debian's choice of system is reasonably sound compared to most or all of the alternatives, because its flaws are relatively minor, technical, and difficult to exploit.

* In principle, you can come up with some weird edge cases where e.g. everybody votes for X and Y over Z (and all other options), but "X is better than Y" receives the same number of votes as "Y is better than X." Then you can't draw an edge between X and Y at all, and you end up with two source vertices (X and Y are effectively tied). As you might expect, the Debian constitution actually does account for this possibility, but it just says that "the elector with the casting vote chooses which of those options wins." The only obvious alternatives would be to pick a winner randomly, or to automatically pick "further discussion" (which basically means "nobody wins, now go find some new candidates and start over").

Possible changes to Debian's decision-making processes

flussence — Mon, 18 Oct 2021 01:50:02 +0000

+1.

I really don't understand Debian sometimes. That's not to say I disagree with how they do things, just that they're doing something way smarter than I can follow. And it seems to be working well for the most part, given how long they've persisted.

Possible changes to Debian's decision-making processes

pollo — Sat, 16 Oct 2021 15:02:20 +0000

Once again, LWN summarizes complex Debian mailing list debates I could not afford to follow. Thanks a lot.

Possible changes to Debian's decision-making processes

AndrewMay — Sat, 16 Oct 2021 12:07:58 +0000

As previously quoted by LWN.net:

It's easy to be part of a community when everyone agrees. It's powerful and delightful to be part of a community when people disagree but the community still works together with respect and mutual support. Creating process that allows myself and others to do this more easily is part of how I enjoy contributing to a community.
— Russ Allbery
(https://lwn.net/ml/debian-vote/8735pof3bj.fsf@hope.eyrie....)

The proposed change is a suggested movement toward bettering a process for the Debian community, which might not reach consensus. However, it does strive to increase "respect and mutual support", which I applaud.

Possible changes to Debian's decision-making processes

anmoch — Fri, 15 Oct 2021 17:50:28 +0000

> you play organized Calvinball

Thank you for such a quotable line. I shall use it at $DAYJOB!

Possible changes to Debian's decision-making processes

mpr22 — Fri, 15 Oct 2021 17:44:10 +0000

I would say rather:

You carefully think about your decision making process in advance, and play organized Calvinball when a weird corner case happens; or you don't carefully think about your decision making process in advance, and play disorganized Calvinball when a perfectly ordinary edge case happens.

Possible changes to Debian's decision-making processes

NYKevin — Fri, 15 Oct 2021 17:41:08 +0000

There are two options: You carefully think about your decision making process in advance, or you play organized Calvinball whenever a weird corner case happens.

Possible changes to Debian's decision-making processes

intelfx — Fri, 15 Oct 2021 16:04:18 +0000

Ah yes, Debian Project is in talks over making a decision to change the decision-making process...

/s, don't take too seriously