The forked elephant in the room

[Don Allen]

On Tuesday, 16 January 2024 at 14:03:20 UTC, Dibyendu Majumdar 
wrote:
> On Tuesday, 16 January 2024 at 13:42:36 UTC, Don Allen wrote:
>> On Tuesday, 16 January 2024 at 10:47:11 UTC, Dukc wrote:
>>> On Tuesday, 16 January 2024 at 10:06:47 UTC, Atila Neves 
>>> wrote:
>>>> but it also can't be the case that the default is to merge 
>>>> PRs unless "there's a reason not to".
>>>
>>> Why not? I get that such PRs are not necessarily net 
>>> positives in purely technical sense, but if there's no reason 
>>> not to merge them they can't be big technical negatives 
>>> either.
>>>
>>> If accepting PRs like that helps to keep people around, then 
>>> considering the morale effect, I'd argue merging such PRs is 
>>> still a net positive.
>>
>> This is an excellent point that I think Walter and the others 
>> who manage this project need to take very seriously. The 
>> technical-social balance of this project is clearly skewed, 
>> the evidence being a long-term pattern of talented people 
>> heading for the exits.
>
> Not at all a good idea. The D team is the gate keeper of the 
> language and have to ensure that each feature integrates with 
> the whole. Saying yes to every feature by default is complete 
> madness. There is a huge cost to every new feature.

I won't speak for Dukc, but I am certainly not advocating "saying 
yes to every feature". That's open-loop stupidity. What *is* 
being advocated is better balance between technical and social 
considerations, which has to be applied on a case-by-case basis.

And there will (or should) be cases that might be rejected solely 
on technical grounds that should be accepted by also considering 
the social cost of rejection. The whole point is to maximize the 
welfare, the forward progress, of the project. These are 
judgement calls that the project leaders need to make and the 
evidence indicates that the way they do it needs adjustment.

Case in point -- the OpenBSD project, for longer than you might 
think was sane, supported the Vax. Why? Because there were some 
people who were making valuable contributions to the project that 
cared about the Vax for some reason and Theo de Raadt, the 
project leader, decided that the cost of indulging them was 
worthwhile, given the benefits of keeping these people around.

>
> Worth watching https://www.youtube.com/watch?v=gXdS3IftP0Y
>
> D is arguably already too full of features because of trying to 
> please everyone.
> As someone argued, its better to focus on quality rather than 
> features at this stage.