I'm a developer. I use main/release/dev for new projects, because it just sounds better and is more intuitive to me honestly. "Master" doesn't make much sense. Like what's so "master" about a "master branch"? It's just the main branch everything gets merged into. It doesn't "control" branches. There's no "master/slave" relationship there. So again, "master" was never really intuitive to me.
Old projects don't get relabeled, they stay master, cause relabeling the main branch could cause potential problems. That's my two cents.
I look at “master” in our repo like you would refer to a master recording or a remaster, or similarly the gold master for when you could say a video game has gone gold.
I don't know what a master recording is. Googled it and it seems to be related to vinyl or something. So yeah, kind of hard for me to wrap my head around that, but definitely an interesting outlook.
Same for databases, master / slave does not really describe the relationship anymore. It's a primary, secondary, control node, read only or something else.
I think that's because in computer science most master/slave nomenclature comes from hardware with a command/control structure (still notable in things like Spark where the namenode/master node controls the data nodes).
GIT just took naming conventions from other existing design patterns (although I should probably look up sources to verify that assumption).
That's where you should use something more like top / bottom /s
I think in this sense, master is more akin to the 'recording' master - The best version of the recording to which others are generated, and all parts merged; no 'slaves' necessarily just the 'master'.
Master can also mean proficiency. If you say you've mastered a trade it doesn't mean you enslaved the trade, you simply have complete knowledge of the trade.
So in that context, the master branch is the complete branch. The branch that other branches stem from because it's the one with code from all the teams. You could branch from another team member's branch but if that branch hasn't merged from master in a while, it won't have all the knowledge (code). When you merge in master you're getting knowledge from elsewhere from the branch that's aware of more things than your branch is: the branch that has mastery of the code, the master branch.
That's not how the terms entered computing though. We always used master in opposition of one or multiple slaves. It implies that one component has control and orders the other one around.
So in a git commit (since they mentioned branches)... What's the slave? Since your the one gatekeeping the word you should know right? How come Git can't be Master in the context they provided when there is no existence of a slave commit?
I'm not sure where you're going with this. I haven't gatekept anything, you can use whatever term you want, that's none of my business. You can happily read my other comment. To me, "master" makes no sense if there are no "slaves". That's why I don't use it. It doesn't make sense to use it. You do you, that's your business.
That’s not how the terms entered computing though. We always used master in opposition of one or multiple slaves.
And yet you said this... Acting like you speak for the entirety of industry, when I bring up one specific facet of our industry that isn't using the term juxtaposition to "slaves"... but rather to other concepts of "master", you now magically change your tune.
Look, I come from the hardware part of the industry and have never seen anyone talk about "master records" in software but always about master devices controlling slave devices. I'll give you that, apparently "master records" are a thing (although I'm curious in what part of the industry). At the same time, it seems so niche and weird to me that there's no point for me to use it. I'll stick with main because it just makes more sense and seems a lot more intuitive to people than to think about master records and what not. You do you, I personally absolutely do not care at all what you go with in your projects.
We're talking about computing here. At least the post does. I guess you could be a QA engineer or something else, but this discussion is mostly a thing with developers.