Re: [RFE] Demilitarize Documentation (was RE: Delivery Status Notification (Failure))

From: Philip Oakley <philipoakley@iee.org>
Date: Feb 19, 2019, 20:16 UTC
Message-ID: <57234e45-69f3-8ff8-b232-eff4985efda5@iee.org>
In-Reply-To: <nycvar.QRO.7.76.6.1902191547510.41@tvgsbejvaqbjf.bet>

Hi all,

On 19/02/2019 14:58, Johannes Schindelin wrote:

Show 32 quoted lines

> Hi Gábor,
>
> On Tue, 19 Feb 2019, SZEDER Gábor wrote:
>
>> On Tue, Feb 19, 2019 at 09:02:43AM +0100, Senol Yazici wrote:
>>> 1. Dictator
>>> Concern: "Bad" connotation.
>> "Benevolent dictator" is a well-established term in open source
>> projects, and it has an inherently good connotation.
> It is a well-established term, alright. Does it have an inherently good
> connotation? No, absolutely not. Every time anybody calls me the BDFL of
> Git for Windows, it annoys me, to say the least.
>
> And yes, when I pull out my generous self, I can give you that the
> *intention* was funny. But to some, it is not funny at all.
>
> Besides, in our field we had pretty established terminology for a long
> time. It was the *architect* who had the final say over what goes in and
> what stays out. And the respective team leaders were responsible for
> respective areas of the code, trusted by the architect.
>
>>> Further, "googling" dictator does not give Linus as a result in (at
>>> least my) search (bubble).
>> https://en.wikipedia.org/wiki/Benevolent_dictator_for_life
>>
>>> Suggestion for substitution: Principal or principal integrator.
>> These are poor substitutions.
> I agree that those are poor substitutions, but shooting down without
> giving better alternatives is a poor way to reply ;-)
>
> Ciao,
> Dscho

I tend to agree that the 'Dictator' perspective is probably a pretty poor choice in modern times, while the other (Lieutenant, Blessed, etc.) word choices are now considered (in much of the millennial society) to be old fashioned, or worse.

However I do caution that we can't be fair all the time. There are a range of impossibility theorems [1,2] regarding making things fair.

Ultimately some one (The Maintainer) must make the discriminatory decision as to what to accept, what to choose, or to reject, or select their own preference.

A probably bigger problem is actually the limited number of actual workflows styles that are recorded (and hence shortage of words for them). For example: patches vs PRs; review process style; even a term for the users 'backup' repo (on GitHub, GitLab, etc) and how it should operate. The mental models here can be hard.

Philip

[1] https://www.youtube.com/watch?v=jIXIuYdnyyk Tutorial: 21 fairness definitions and their politics, Published on Mar 1, 2018

[2] Arvind Narayanan: Associate professor of computer science at Princeton <https://www.youtube.com/user/33BitsOfEntropy>