[FFmpeg-devel] [RFC] clarifying the TC conflict of interest rule

[Marton Balint]

On Tue, 20 Feb 2024, Anton Khirnov wrote:

> Quoting Marton Balint (2024-02-20 10:12:34)
>> We have no means to prove financial interest, because it is not public.
>
> We also have no means to prove that committee members are acting in the
> project's interest.
>
> E.g. if I had no qualms about being dishonest, I could always ask a
> friend to object to controversial patches in my place, so I wouldn't
> lose my vote, and nobody could prove it.
>
> In the end some things have to be taken on trust.

My concern is bad mouthing others based on assumed financial interest and 
endless discussion if that interest is "serious" or not. If your payjob 
uses ffmpeg, or if you ever want money for some ffmpeg related work, that 
is a financial interest right there.

If somebody feels that voting would not be fair, he can always abstain. 
I'd rather keep that fully trust based, to avoid rule interpretation 
wars and discussions about assumed interests.

An interest is not inherently bad, selfish contributions (financial 
reasons or not) is a huge factor in open source.

>
>> For practical reasons, using patch authorship is better. Or maybe a more
>> general solution against bias is somewhat increasing the number of people
>> in the TC, and removing this rule alltogether.
>
> I woould be concerned about making the TC too slow and unwieldy, it
> already takes a lot of effort to push any decisions through. Keep in
> mind that during all of its existence it only ever made two decisions,
> and one of them spent over a year in limbo.

So with 7 people, it would have been two years? :)

Have the TC meet weekly if there is an agenda, and have votes after 
meeting. With more people it is not that big of a deal if somebody 
cannot attend. And have a rule in place to resolve ties. It can be as 
simple as to accept the proposal of the party who raised the issue to the 
TC.

Regards,
Marton