I think this is a misunderstanding. The way I read the RFC, you need to pick a hash function as was suggested by Alex. The use of the RFC is mandated by our bylaws, I disagree it is complex, and so far you are the only one arguing against it, without any clear benefit to your alternative suggestion. This resulted in holding up the completion of the election, as your concerns prevented us from concluding a procedure in the meeting, and using the Friday draw.

In the interest of completion I'd take rough consensus as our guideline, as is our usual approach. Also note that both candidates expressed being fine with the proposal. So unless I see anyone else arguing against Alex' proposal, I think Alex should just go ahead as proposed, and prevent further bikeshedding.

--
ralphm


On 22 November 2025 22:50:24 CET, Travis Burtrum <travis@burtrum.org> wrote:
Hi Alex,

As I said in the meeting I see no reason to do some complicated procedure including a hash before doing 'mod 2' and picking the winner when we could just do 'mod 2' on the value of the first ball for a fair tie break.

The only reason anyone brought up so far is "we made this mistake in the past already" which seems like a bad reason to do it again.

Thanks,
Travis


On November 20, 2025 7:15:11 PM EST, Alexander Gnauck <gnauck@gmail.com> wrote:
Hello,

you can find the meeting minutes of our annual board and council election meeting here:
https://wiki.xmpp.org/web/Meeting-Minutes-2025-11-20

All Council candidates were accepted. The following individuals will form the XSF council for the 2025/2026 term:

* Dan Caseley
* Daniel Gultsch
* Jérôme Poisson
* Stephen Paul Weber
* Marvin Wißfeld

For the board candidates we have a tie on the 5th position between Adrien and Arne. Our bylaws state that the fifth candidate will be chosen with:

RFC 3797: Publicly Verifiable Nominations Committee (NomCom) Random Selection

The currently elected individuals for position 1-4 are:

* Guus der Kinderen
* Mickaël Rémond
* Ralph Meijer
* Florian Schmaus

I am suggesting that we use similar procedure to what we used in 2017 when we had our last tie. The solution was proposed by Dave Cridland. Its compliant with our bylaws.

The proposal is the following:

1) As random input source, I propose using the "Main Numbers" from the
Euro Millions draw of next Tuesday (2025-11-25), as announced here:
https://www.euro-millions.com/results

2) Each number will be arranged in ascending order, separated by dots,
and terminated by "./".

3) The resultant string will be hashed according to SHA-256.

4) The tied candidates will be arranged into alphabetic order (note,
in this case this may be by first or last name, it makes no
difference), to produce:

* Adrien Bourmault
* Arne-Bruen Vogelsang

5) These will be numbered from 0. Adrien is 0, and Arne by 1.

6) The decimal representation of the last byte in the hash, modulo the number of candidates will then be used to select the candidate.

As example, if we take these numbers:

02 - 10 - 14 - 28 - 31

We'd form the string with the numbers ordered of "02.10.14.28.31./"
Which we can hash with:

sha256('02.10.14.28.31./').hexdigest()

Producing a hash (in hex) of
'ad2edee63a88d7b4b6109944b5222a6cce1c3719ac4d41acd0dbf8287465ace0'

The last byte in the hash 'e0' is 224 in dec.
224 mod 2 = 0

This would result in candidate 0 being selected (Adrien).

I have created a small script to create the results with this algo.

It is on GitHub here:
https://gist.github.com/agnauck/8996415c1a91e5b6e1d27cad0367c85b

and you can run it online here:
https://sharplab.io/#gist:8996415c1a91e5b6e1d27cad0367c85b

Alex