[FFmpeg-devel] [PATCH 1/2] avcodec/s302m: enable non-PCM decoding

[Michael Niedermayer]

On Mon, Feb 19, 2024 at 10:37:15PM +0100, Anton Khirnov wrote:
> Quoting Michael Niedermayer (2024-02-18 23:34:39)
[...]
> > > > But I think it is reasonable that parties of a disagreement cannot be
> > > > the judge of the disagreement.
> > > 
> > > Why not? This is one of those truthy-sounding statements that does not
> > > actually hold up to scrutiny.
> > 
> > * A disagreement implies that there are 2 parties
> > * And we assume here that what one party wants is better for FFmpeg than what the other wants.
> > * The TC needs to find out which partys choice is better or suggest a 3rd choice.
> > * If one but not the other party is a member of the TC then this decission becomes biased if that member votes
> 
> This example is flawed in at least two following ways:
> 
> First, you keep comparing TC members to judges in a legal system. As I
> said above - in a paragraph you ignored - I do not think that is a
> meaningful comparison. We have no law, TC members are not judges and
> decide based on their experience and opinions.

well, TC members make decissions, lets call them decission makers then

lets see how that would look
so if we had judges and law we seem to agree that a decission maker cannot
be a party to the very disagreement she decides on.

lets take away the law
so a courtroom with 5 judges, deciding on a persons fate, no more law everyone
can choose as they prefer.
The judge accusing another man of something, can now be a judge of that case ?

No, the removial of "the law" isnt making the common sense rule any less common
sense

so lets now not call teh judges "judges" anymore, and lets pick them identically
to how TC members are picked
Does this change anything ?

The decission maker accusing another man of something, can now be a decission maker of that case ?

No, the issue remains. A party to a disagreement cannot be decission maker in
the disagreement. There is bias and if the goal is a optimal decission
we want no bias.


> 
> > Imagine a judge kills someone and judges himself innocent afterwards in a panel of 5 judges
> 
> Second, in this example the judge in question has two roles in the
> situation: that of a criminal who wants to avoid being found guilty and that
> of a judge who is supposed to find criminals guilty. The interests of
> these roles are in conflict, hence we have a conflict of interest.
> 

> That does not translate to the situation we are actually dealing with.
> My interests in my role as a patch reviewer and as a TC member are
> exactly the same. There is thus no conflict of interest.

i disagree

A TC member who wants to block a patch and wants to decide if a patch should be
blocked is in the same situation as

a Judge who wants to sue someone and wants to judge that someone.

Again, the judge being a party to a lawsuite cannot be judge in that lawsuite
Similarly a TC member initiating a conflict cannot judge in that same conflict
The TC member surely does the same thing in both cases, he wants something
like blocking a patch.
Similarly the judge also wants to whatever is the goal of his lawsuite

And there are in fact more problems
Consider this:
Normally we have PartyA and PartyB in a conflict and then we have 5 TC members
who look into that, both parties can argue their case in front of the TC (publically)
and the TC then discusses and make a decission, possibly asking more people an so on

But now the case here changes we have PartyA and PartyTC
Theres a disagreement between a developer and a TC member

The TC member is part of the TC and discusses with the other 4 members, he will
support himself in all votes and potential private arguments.
PartyA here has substantially worse chance to win even if PartyA is correct and
has the better solution.

Iam sorry but i insist that the TC member in this case cannot act both as a party
to teh disagreement and as a member of the TC in the same disagreement.

It is completely unfair to partyA in this example above, they are not on equal ground.

thx

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

I have never wished to cater to the crowd; for what I know they do not
approve, and what they approve I do not know. -- Epicurus
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20240220/07f12729/attachment.sig>