top | item 39872793

(no title)

realjhol | 1 year ago

Then main branch story was utterly ludicrous. Imagine thinking that you were helping the world in any way by changing the name of a git branch

discuss

order

int_19h|1 year ago

Even if you don't care about any of the political reasons, it's still two letters shorter, and it's easier for novices to grok.

IshKebab|1 year ago

If those were the only benefits nobody would think it was remotely sane. The extra work supporting `main` and dealing with the confusion far outweighs that.

It would be like renaming `creat` to `create`. Yeah it should have been `create` in the first place but it would be completely insane to change it now.

voidnap|1 year ago

Right, fewer letters equal better! Very professional.

Honestly though, it doesn't matter to me what you or I name branches or what test images we use. I'm going to objectify women and own slaves in any case.

Lio|1 year ago

What makes that even more ironic are the number of people involved in that spat that were probably happily wearing cheap cotton products of dubious origin at the time.

Actual modern slavery still exists and is sadly wide spread.

MaKey|1 year ago

Kubernetes changing terminology from master to control-plane actually caused a reddit outage. It's ridiculous.

pgeorgi|1 year ago

There have been issues with git's master->main transition as well, once automated systems came into play that expected to find a master branch. This is also my main complaint about that particular episode: Too much "somebody needs to do something" mandates, and too little "how to do it properly" (which would provide benefits for other situations as well.) When the two parts of the activity are split between two different groups, there's no incentive, either: The mandate group checks a box and walks away happily, having reached their OKR. The implementation group doesn't get the devops time to implement proper aliasing and what-have-you, so they just wing it until everything works again.

That said, it's mostly water under the bridge right now, and it isn't applicable at all for the "reference image for computer graphics papers" situation unless somebody starts rewriting all the old papers to reference a different image:

A "somebody needs to do something" mandate would likely lead to new versions of the old papers in which the image is removed without replacement. The "do it right" solution would lead to people replicating the research with different material, which might not even be for the worst - but I see no chance that will happen at scale.