Hacker News Reader

Comment by Pfhreak | Hacker News Reader

If you look at the results they link, the picture becomes clearer.

They proposed 8 options. 2 directly called for resignations at FSF, 2 directly supported RMS, one said to stop working with FSF, one said to call for governance improvements at FSF, one was no comment, and one was discuss further.

The two options in support of RMS were not majority held opinions, and were held by ~1/6-1/3 of voters.

No comment beat "call for the removal of RMS" by only 8 votes.

For me, the takeaway here is that while Debian will issue no comment, 2/3+ of voters in Debian supported calling for RMSs removal. It's clear which way the tide is moving, even if Debian says nothing.

Source data: https://vote.debian.org/~secretary/gr_rms/results.txt

Edit: My 2/3+ estimate is a napkin math overestimate, and the number is probably closer to the 50-60% other comments have suggested below.

View on HN

How do you get to 2/3rds? I am not familiar with Debian's voting system, but I get just under 1/2 in favor publically supporting removal.

I define "support" as preferring an option over "Further Discussion".

According to [0], there were 420 unique voters. Using 420 as the denominator, and collumn 8 [1] as the numerator, we get the following approval voting results:

1) Call for the FSF board removal, as in rms-open-letter.github.io - 203/420 = 48%

2) Call for Stallman's resignation from all FSF bodies - 222/42 = 53%

3) Discourage collaboration with the FSF while Stallman is in a leading position - 19/420 - 52%

4) Call on the FSF to further its governance processes - 221/42 = 53%

5) Support Stallman's reinstatement, as in rms-support-letter.github.io - 52/420 = 12%

6) Denounce the witch-hunt against RMS and the FSF - 84/420 = 20%

7) Debian will not issue a public statement on this issue - 277/420 = 65%

8) Further Discussion - N/A

It does not look like individual votes have been released yet. If they were, I would like to see how accurate my assumption that no one "approves" of both his presence and absence from the board. The sum of approval for both 1 and 5 is 255 = 61%, so it seems like most of the no statement voters actively disprove of a statement in either direction (instead of simply preferring no statement)

[0] https://vote.debian.org/~secretary/gr_rms/

[1] https://vote.debian.org/~secretary/gr_rms/results.txt

I may be misinterpreting or misapplying how they do a "passes majority" check, which seems to indicate 222 for and 166 against for option two.

So it might be more correct to say ~3/5s?

https://en.wikipedia.org/wiki/Condorcet_method

You can look at the Majority results if you want to say how many people think an option is acceptable. That ratio is also in the graph. For option 1, 203 people indicated that it's acceptable, 186 said it was unacceptable, leaving 31 that didn't say anything about it. The 203/186=1.09 is in the graph. You could also see that as 52% find that option acceptable.

> Debian will issue no comment, 2/3+ of voters in Debian supported calling for RMSs removal. It's clear which way the tide is moving, even if Debian says nothing.

FWIW, some people didn’t want to vote because the results would be publicly with names.

So the result of a secret vote might have been different.

They should have gated this vote with a yes/no vote on whether to vote.

No, it simply shouldn’t have been public. The DPL elections aren’t public either.

correct, but the debian rules on this issue state that only the elections are not public. there are no provisions for private votes on resolutions.

this may change in the future if enough people find that (optionally) private votes on resolutions are a good idea.

The "Further discussion" option is meant to allow for that. Note that the options calling for a statement in support of RMS were even less preferred than "Further discussion", which is quite weird given the context so GP may well have a point.

Why do you think it is quite weird? Historically, is the "further discussion" always at the bottom?

edit: I found those, so it's not that weird, but not very common: Init systems https://www.debian.org/vote/2019/vote_002 ; Delaying Lenny https://www.debian.org/vote/2008/vote_003 ; Change to membership procedures https://www.debian.org/vote/2008/vote_002

the more controversial or sensitive a topic, the more likely it is that people will believe that more discussion is needed before a decision is reached, so i don't find that at all surprising.

No, they should have used a standard process for general resolutions and not special cased it. Which is what they did. They have a constitution and a democratic process that requires proposals, seconding, a period of open discussion, and then a vote.

They agreed on their voting system in advance exactly to handle cases where the vote might be contentious.

Your hypothesis is that people voted opposite of their beliefs? Or that they chose not to vote at all?

Both of those feel like possibilities, but probably not in any significant number. In absence of evidence to the contrary, I'd say it's reasonable to look at the results of a vote as evidence of how people actually feel.

Edit: Vote numbers seem consistent with other votes in Debian's history, so it does not appear that a large number of people elected not to vote.

I'm surprised Debian didn't invent a way to do secure secret voting by email already (and by some of the technical comments on the results page I kind of though it was secret)

Debian knows how to do secret voting. In fact, Debian's latest election of their Project Leader which also ended last night was secret.

It looks like it would need a General Resolution to allow secret voting for future General Resolutions.

To clarify, 2/3 supported calling for removal over calling for RMS to stay, or "Further discussion". The option to not comment was clearly preferred over others, it was the only choice in the Schwarz set. There is no ambiguity at all about the outcome.

I didn't imply there is ambiguity in the result. But there is a difference between "what we agree on as a whole community" and what the majority opinion is.

The community agreed on no comment, full stop.

A significant preference was also shown for removal over reinstatement.

> 2/3+ of voters in Debian supported calling for RMSs removal

2/3+ is overestimate.

About 52% (228/439) preferred one option from 1-4 (anti-RMS) better than neutral options (7-8).

If we ignore winning option 7, then 63% preferred 1-4 to default option 8, while 20% preferred 5-6 (pro-RMS) to option 8.

Yeah, it was definitely napkin math. I'll edit my comment.

What in the hell is with that tabulation of data?

You don't even have raw vote counts. You have that weird floating relative representation going on, and no formula is given for the totals, or I'm not aware the procedures are that line that out. Example:

2 dimensional matrix Value at (x, y) coordinates should be read as option X received <value> more votes than option Y. The diagonal is blank since it degenerates in this representation.

Row 1 column 2 would assert that option 1 beat option 2 by 113 votes.

Row 2, column 1 however asserts the opposite, that option 2 beat option 1 by 137. This is reading is articulated explicitly in those results.

I see no way that isn't an indication of contradiction unless their tallying method is so bloody exotic I can't even imagine. Nor can I find the where they define their dividend for the majority calculation, but I'm assuming there is a tally of nay/abstain votes somewhere?

What am I missing?

Voting procedure is documented here [0] section A.6

I agree that it would be nice to include the total number of ballots in the final results, but that isn't really relevant to the results.

Each voter submits a ranked choice vote ballot. The table indicates that 113 ballots prefer option 1 over option 2, whicl 137 ballot prefer option 2 over option 1.

The majority calculation is a bit weird. All Debian votes include default option, in this case that is option 8, "Further Discussion". Preferring an option over the default is seen as a general "approval" vote, while prefering the default is considered "disapproval". The "majority" votes looks at the ratio between approval and disapproval. For instance, option 1 had 203 votes to apporove and 186 votes to disaprove, giving it a ratio of 203/186 = 1.09

Voters are not required to vote on all options. Excluding private votes, they release every individual vote once voting is closed, so you can see how many abstensions an indidual option received; but it appears that this has not been released yet.

[0] https://www.debian.org/devel/constitution

That'd be it. There was an entire dimension going on there I was unaware of the ranked choice. I was thinking FPTP. Thank you for the reference.

So there is no general 'nay' vote on a proposal except to vote it at a lower rank than the default option? I think I get it now.

I'm actually kind of glad most people don't want to make any statement at all. It isn't really something tht should matter. Beyond my vocal opposition to lending any credibility to the original smear with regards to RMS, I'm kind of in the camp we should all just move on. Most of FSF's insiders that have a problem with him have more than enough points actually relevant to his missing opportunities to make a worthwhile push at ousting him without relying on controversy and lies, but they'd also have to defend it from an equal playing field then. LLVM, GPL issues, etc... Though even looking into the LLVM controversy, I can still see the wisdom in not accepting it and keeping GCC monolithic after having come across a few oldtimers who were there back when proprietary everything was the norm.

If someone is to be dismissed, let it be on technical merits. Not some tabloid smear.

I think the language in results.txt is a little confusing. What I think it's saying is:

113 people preferred option 1 over option 2

137 [different] people preferred option 2 over option 1

and presumably the balance of the 420 didn't express a preference. Or at least that's how I read it.

> 2/3+ of voters in Debian supported calling for RMSs removal

How did you come to that number?

Look at the row for option 2 in the data table, it will show you how many people preferred that option over other options. Roughly two thirds preferred removing RMS over "we should talk about this further", for example.

Row for option 2 is:

Option 2 137 168 165 239 227 199 222

222 voted for option 2 over "we should talk about this further" (default position), which is ~53% (222/420).

Was it an anonymous vote?

No. Do you have evidence that anonymity would have mattered?

The vote totals seem in line with historical votes.

It's a politically controversial topic, and preference falsification is a well-documented sociological phenomenon in such circumstances.

If you want to know what people really think without the possibile influence of preference falsification, it needs to be an anonymous poll.

> If you want to know what people really think without the possibile influence of preference falsification, it needs to be an anonymous poll.

Which is exactly why we not only allow, but actually enforce it in elections. If your vote could have a strong impact on e.g. your future employment, you wouldn't be free to vote for what you think is best, you'd be compelled to vote for what you think is perceived best.

Except that's not Debian's agreed-to voting process.

I'd also point out that secret ballot is not universal in the US. Certainly the votes of elected officials are not secret. But neither are open votes at, say, a town hall meeting.

> Except that's not Debian's agreed-to voting process.

I'm not suggesting they'd have to do that, I'm saying that there's a very good reason why we usually do that. As organizations become more and more political, I expect that to hit harder and for more of them to adopt secret voting.

> Certainly the votes of elected officials are not secret.

And I find that reasonable. You want to know whether the person you voted for actually does what they said. In Germany, parliament votes are public, but important votes are secret.

For the less important votes, they'll just ask for raised hands and note who won, but not who voted in what way (we now have TV channels film the votes and then use interns to determine who voted what).

For important votes, they're done secretly: you can tell who voted, but not what they voted for. There's no way to hold them accountable, the system is working as intended.

Sure, but do you have evidence it would have moved the needle in the vote. There are many, many well documented phenomena that influence every vote. While they may influence a small number of voters, I've yet to see any evidence that these results are not the genuinely held belief of most folks.

No, I don't have specific evidence, and I won't have it until we see an anonymous poll.

I do believe I have a very plausible hypothesis, though, and I find it sufficiently plausible that I'm going to reserve judgement on the true beliefs of people until we see that anonymous poll.

it is conceivable that voters who are pro-rms voted 'no statement' or 'further discussion' instead

How is it political? What politics are even involved? This very much is a personal issue with Stallman. His objectionable behavior isn’t tied to political beliefs, it’s tied to an individual’s moral and social failings.

Fine, it's just a plain-old controversial topic, then. The same sociological pressures that can lead to preference falsification are in play as with any contentious issue.

Are you asking for evidence that secret ballots prevent voter intimidation? Or evidence that people are harassed online for their stance on this specific issue?

>our comparison of survey results with official election truth indicates that the use of a so-called secret ballot yields more accurate, valid self-reports of voting on socially sensitive ballot issues, such as supporting taxes for elderly services.

https://academic.oup.com/poq/article-abstract/59/4/568/19401...

>regression analysis shows that social forces appear to have a greater effect on vote choices among people who doubt the formal secrecy of the ballot.

https://www.jstor.org/stable/23526132

Neither. I'm asking for evidence that this phenomena would have materially moved the result. I'm willing to believe that 5% of people cast a vote here contrary to their opinions (though that feels high). Even so, the results likely don't shift materially.

> It's clear which way the tide is moving

It's only clear that the way this is measured is heavily biased: 4 options out of 8 are against current FSF/RMS' statu quo. Only 2 are for supporting the statu quo. And 1 is for indifference, and the last one for more discussion needed.

If you overcharge the negative choices, it's kind of given most of the answers will fall in them. Survey design 101.

For reference:

- Option 1 "Call for the FSF board removal, as in rms-open-letter.github.io"

- Option 2 "Call for Stallman's resignation from all FSF bodies"

- Option 3 "Discourage collaboration with the FSF while Stallman is in a leading position"

- Option 4 "Call on the FSF to further its governance processes"

- Option 5 "Support Stallman's reinstatement, as in rms-support-letter.github.io"

- Option 6 "Denounce the witch-hunt against RMS and the FSF"

- Option 7 "Debian will not issue a public statement on this issue"

- Option 8 "Further Discussion"

> If you overcharge the negative choices, it's kind of given most of the answers will fall in them

That's not how it works. The voters are asked to rank all options according to their preference, and a winner is chosen according to an extended Condorcet criterion. Adding similar options doesn't affect the outcome.

You should read up on how they conduct votes. Every voter essentially gets a chance to rank every option against every other option. So the number of options should not impact the vote totals.

You forget human perception. If I expose someone to 10 negative statements and only 2 positive ones, you believe that this does not affect their overall perception of the issue?

I'm sure it does, but I don't believe it does materially in this case. People are going to have a perception of other terms, e.g. RMS, as well. While yes, what you describe may have influenced an emotional response, I see no evidence that would have changed the outcome.

You do realize that the community needs to propose and sponsor these options, right?

Fewer people being willing to propose a pro-RMS option is an indication of which way the tide is moving, even if you still want to argue about why this might be the case.

it only takes a few people to sponsor options, and plenty of people have sponsored options that they would not vote for because they believed that these options should be available.

That's exactly my point. If sponsoring options is done by the community, and is relatively easy, then fewer proposed options in one category does suggest that this category is unpopular with the community.

If there are several variations of anti-RMS proposals, it suggests that enough people care about the specifics of how the rebuke is worded, etc., but that generally there is enthusiasm about an anti-RMS stance.

If there are fewer pro-RMS proposals, then either they were proposed by people who don't support that option, but are putting it forward because they think it should exist as an option (as you suggest), or because there are fewer people who are enthusiastic enough to care about the minutia of how a support message is to be worded. Either way, that seems to me like less enthusiasm around an anti-RMS proposal.

new options are only proposed when existing options do not represent all available opinions.

similar proposals can also be merged, and only options that make sense are being sponsored. and they are sponsored also by people who oppose that option because those people believe that the options should be balanced.

the existing options list all sensible outcomes. i can't think of any other pro- or anti-RMS option that would even make sense.

based on that i can't see how you want to extract any data on popularity

Reply on news.ycombinator.com