Re: Abusive cancels from miakibot (bis)

llpminibot avait énoncé :
> Dans son message précédent, pehache a écrit :
>> Following a previous discussion on that I just want to remind the 
>> (approximate) recent history.
>>
>> A few months ago (in 2021, maybe 2020, I don't remember exactly) Olivier 
>> Miakinen (O.M.) wrote a cancel bot named "miakibot" to clean up the fr.* 
>> hierarchy from spam etc, as many cancels were done manually before that.
>>
>> At the begining all was OK, nobody was complaining, and most people were 
>> even praising it. There was an old FAQ on fr.* about legit cancels and the 
>> bot was essentially cancelling according to this FAQ.
>>
>> At the same time a usenaut, whose latest incarnation is known as "Néo 
>> L'élu", who was posting tons of articles, always with the same kind of 
>> subject (about religion, atheism...), with crossposts and no follow-up. The 
>> guy was indeed a pain in the ass, but most of time he was not violating any 
>> formal rule of fr.* that could justify the cancellation of his posts.
>>
>> To try limiting the amount of messages from Néo we thought about updating 
>> the FAQ to forbid crossposts without fu2 between fr.* and another hierarchy 
>> (as a matter of fact Néo was crossposting on francom.*). We took this 
>> opportunity to completely rewrite the FAQ, as we felt it was outdated. 
>> There are has been a discussion on that last automn on the group 
>> fr.usenet.abuse.d, and we reached a consensus on a new FAQ.
>>
>> However, soon after that O.M. decided to cancel ALL posts from Néo, even 
>> the ones that were not violating the rules described in the FAQ. The reason 
>> he gave was that Néo was flooding fr.* anyway. This is a kind of indivual 
>> UDP, and several people expressed that such an extreme decision should not 
>> be taken by a single person, and that a *consensus* should be obtained on 
>> fr.usenet.abuse.d.
>>
>> O.M. then opened a discussion about that, and about some other individual 
>> UDPs applied to some other spammers/flooders. He obtained large majorities 
>> for almost everything, except for Néo. There was just a small majority (12 
>> out of 22) in favour of the UDP on Néo.
>>
>> One can endlessly argue about "what is a consensus". A common and pragmatic 
>> definition is "a large majority". Still one can argue about "how large ?", 
>> but what is sure is that a 12 out of 22 is NOT a large majority, and hence 
>> is NOT a consensus.
>>
>> By the meantime the miakibot was still cancelling ALL articles from Néo, 
>> and O.M. didn't want to change anything. After a while, however, and 
>> because of many protests, he stopped cancelling Néo.
>>
>> But in march 2022 he resumed cancelling Néo. Just, instead of cancelling 
>> all articles, he allowed him to post on a few selected groups 
>> (fr.sci.zetetique and the groups related to the Apple Mac), and without any 
>> crosspost. All other articles were cancelled. And again it was his own 
>> decision, taken without any consensus.
>>
>> Somehow acknowleging the problem, O.M. used a different "From" for the 
>> cancel articles on Néo: "miakibot2 <[email protected]>" instead of 
>> "miakibot <[email protected]>".
>>
>> Soon after he started to cancel the answers to cancelled articles. This 
>> would be reasonnable if all the cancels were legit in the first place, 
>> which is not the case. Moreover no consensus has been obtained for 
>> cancelling the answers to cancelled articles. The "From" for these cancel 
>> articles is "miakibot3 <[email protected]>"
>>
>> So at the moment we have :
>>
>> - miakibot: legit cancel articles, supported by the FAQ on legit cancels or 
>> by a consensus
>>
>> - miakibot2: abusive cancel articles, not supported by the FAQ or by a 
>> consensus
>>
>> - miakibot3: abusive cancel articles, not supported by the FAQ or by a 
>> consensus
>>
>> At least, cancel articles from miakibot2 and miakibot3 should not be 
>> honored.
>>
>> Cancel articles from miakibot are legit at the moment, but one can wonder 
>> if O.M. can be trusted on the long term. At first he looks opened to 
>> discussions, but he is actually always deciding on his own, without 
>> bothering about a consensus, and he is convinced that his point of view is 
>> always the right one (if you do not agree with him, then you're just a 
>> troll). Also it is worth mentioning a massive abusive cancellation last 
>> year, when he *retroactively* cancelled thousands of articles of a single 
>> usenaut. Later on, he admitted that he did that because he had been upset 
>> by this usenaut. Giving the atomic bomb to someone who is emotionnaly not 
>> that stable is maybe not a good idea...
>>
>> Note: there is currently no more "miakibot2" cancel article, but the reason 
>> is just that "Néo" has disappeared from fr.* (his Eternal September account 
>> has apparently been suppressed because of violations of conditions of 
>> service"). He may come back, or O.M. may decide one day or another to put a 
>> UDP on someone else... So blocking miakibot2 (and miakibot3) is still 
>> appropriate.
>>
>> Note 2: all of this also raises some questions about the alphanet.ch 
>> server. Not only O.M. uses this server to send the abusive cancel articles, 
>> but the Alphanet newsmaster is indeed fully supporting him.
>
> I totally agree with you.
>
> The miakibot must be stopped.
> As you have very well reminded, we cannot trust its author.

A small example of a "test" message cancelled by the miakibot3.
No matter the content of the message, it will be cancelled.
Message-ID: 
<cancel.e4kn9h1t94m7osqg0k7eqho3udiktdchq4@consensus-omnium>