Hacker News Reader

Comment by cphajduk | Hacker News Reader

To me, "master" and "slave" are historic terms used throughout electrical, software, even entertainment industries.

Eliminating words from a vocabulary is very 1984-like. Those words have a deep historical meaning, allowing ourselves to just "remove them" is akin to forgetting and ignoring the dark past of slavery, rather than remembering and acknowledging it (with the hopes it will never happen again).

Saying that it helps change habits (in my opinion) is analogous to saying that preventing kids from playing violent video games will reduce mass shootings (there is evidence it does not). I disagree with your premise that this pushes us to change habits and is only a mechanism to be ashamed of our shared (and dark) history. Lest we forget.

View on HN

Github isn't eliminating any words from anyone's vocabulary.

This thread has demonstrated that plenty of people are committed (har har) to calling their repository's je ne sais quoi branch `master`.

While I'm with you that I don't understand how this will move the needle on racial equity, I'm uncomfortable with how visceral of a reaction a group of technology professionals is having to what is essentially a library changing a default value.

Like, vocabulary changes all the time. Technology changes even more frequently. Why y'all so scared to use a different label?

When you write software, you should only change a default value for a good reason. This was...not that.

I agree that the outrage can sometimes seem out of proportion to the change itself, but I can also understand why people who write software in general would be offended by the silliness of the whole episode.

Has Github stated a reason?

I read the announcement from Oct. 1 [1] and it doesn't have any explanation outside of a link to a Software Freedom Conservancy [2] (the folks now maintaining Git)

A lot of people here are assuming virtue signaling, but it could just as easily be "a majority number of our staff was behind this change". Unless GitHub has stated the why somewhere (I spent ~5m googling to no avail) we simply don't know.

[1]: https://github.blog/changelog/2020-10-01-the-default-branch-... [2]: https://sfconservancy.org/news/2020/jun/23/gitbranchname/

> Why y'all so scared to use a different label?

Because you just created a massive amount of tech debt that needs to be addressed in the here and now without convincing people that creating this tech debt was worth creating in the first place and when there is a lot of other tech debt that actually matters that still hasn't been paid off.

Like someone else said here: "On one hand, here I am trying to get work done and on the other hand you have these people actively slowing me down. These people are my enemy"

This change only affects newly created repos, how does it create tech debt? I suppose some tooling may need to be updated, but if your tools are to brittle to support a different branch name.. sheesh

I would posit to your quotee that they're being phenomenally self-centered.

Github has been mum about the why behind this change, but I'd bet my hat it wasn't because they wanted to actively slow down't their users.

At my organization, we are being pressured to change existing repos to use "main" with the implication that we are racist if we do not. But even if we leave existing repos alone, now we all have to remember which repos use master and which ones use main. We tend to have people working across many different repos, so it's a headache waiting to happen either way.

> This change only affects newly created repos, how does it create tech debt? I suppose some tooling may need to be updated, but if your tools are to brittle to support a different branch name.. sheesh

This is overly dismissive. Build pipelines that interact with bespoke branches now need dynamism for backwards compatibility; a value that was previously static is now changeable. That doesn't really qualify as brittle to me; that any value in a codebase must be changeable is a ridiculous requirement from a codebase.

Don't forget documentation. Massive amounts of tutorials and FAQs will now be more confusing to newbies.

Can you give me a real example?

I work with build systems in my day-to-day, and I can't remember the last time I worked with something that didn't support dynamic branch names but did support git

But my experience is obviously skewed by where I work.

I'm specifically thinking of git-flow (https://nvie.com/posts/a-successful-git-branching-model/); every build system I've interacted with has been some flavor of this. The crux here is that there is a single branch that deploys occur from. Not uncommonly, this is the default branch.

Sure, I defo have opinions on git-flow.

But with every build system I work with (which are: Jenkins, Concourse, Github Actions, and Gitlab CI) you can make any branch you want the branch-to-build-on.

I don't mean to say that it's not totally fixable. Up until this change, it was a reasonable assumption for any org to make, that the default branch will be the same for all projects. Now, either the default branch on any new repo must be manually set to the old default, or the build system must be updated to handle non homogenous default branches.

> This change only affects newly created repos, how does it create tech debt?

Every book and piece of documentation on git is now obsolete. People learning git will now hit a wall trying to do very basic things.

Reply on news.ycombinator.com