In brief,

On Thu, Apr 2, 2026, at 7:16 PM, Michael Niedermayer via ffmpeg-devel wrote:
> Instead of applying that apparent consensus, or first announcing that a 
> vote would be started, Niklas opened a vote.

It was announced over two different channels:

1. On the mailing list: 
https://lists.ffmpeg.org/archives/list/[email protected]/message/ZQVTFLYU7ZCMRLN46P2UHWXSPNDBWSAV/
2. On the issue tracker: 
https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22546#issuecomment-32978

> 2. The set of choices was defined by Niklas, while others had very 
> little time to comment. I explicitly said that we needed more time and 
> more discussion.

This is not true; most of the options were proposed by others, not me.

Also, in two weeks of discussion, nobody so far has proposed additional 
options. I asked you about five times what options you think are missing, and 
received no reply.

> 3. This is a vote about how to integrate checkasm, started by one of 
> the main checkasm authors, with checkasm authors also participating in 
> the vote. At the very least, that raises a conflict-of-interest 
> question that should be addressed explicitly before relying on such a 
> vote.

This is a vote about what the FFmpeg community prefers, and mainly affects the 
FFmpeg project. The impact on the checkasm project is minimal.

> A vote should be a last resort after discussion has failed, not a 
> substitute for discussion.
>
> [...]
>
> In my view, this vote should either not happen at all, because we 
> seemed close to consensus, or it should be restarted and done properly.

For context, 6 different developers have expressed mutually incompatible 
preferences, and at least three people have agreed that a vote is the best 
solution.

Whether or not you believe the vote to be legitimate, I will resubmit the PR 
based on the results of this vote.
_______________________________________________
ffmpeg-devel mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to