BerandaComputers and TechnologyWhy renaming Git’s master branch is a terrible idea

Why renaming Git’s master branch is a terrible idea

Back in May (in the inauspicious year of 2020) a thread in the Git mailing list with the tile of “rename offensive terminology (master)” was started, it lasted for more than a month, and after hundreds of replies, no clear ground was gained. The project took the path of least resistance (as you do), and the final patch to do the actual rename was sent today (November).

First things first. I’ve been a user of Git since 2005 (before 1.0), and a contributor since 2009, but I stopped being active, and only recently started to follow the mailing list again, which is why I missed the big discussion, but just today read the whole enchilada, and now I’m up-to-date.

The discussion revolved around five subjects:

  1. Adding a new configuration (init.defaultbranch)
  2. Should the name of the master branch be changed?
  3. Best alternative name for the master branch
  4. Culture war
  5. The impact to users

I already sent my objection, and my rationale as to why I think the most important point–the impact to users was–not discussed enough, and in fact barely touched.

In my opinion the whole discussion was a mess of smoke screen after smoke screen and it never touched the only really important point: users. I’m going to tackle each subject separately, leaving the most important one at the end, but first I would like to address some of the most obvious fallacies people left at the table.

The obvious fallacies

The most obvious fallacy is “others are doing it”. Does it have to be said? Just because all your friends are jumping off a cliff doesn’t mean you should too. Yes, other projects are doing it, that doesn’t mean they don’t have bad reasons for it. This is the bandwagon fallacy (argumentum ad populum).

The second one comes straight out of the title “offensive terminology”. This is a rhetorical technique called loaded language; “what kind of person has to deny beating his wife?”, or “why do you object to the USA bringing democracy to Iraq?”. Before the debate even begins you have already poisoned the well (another fallacy), and now it’s an uphill battle for your opponents (if they don’t notice what you are doing). It’s trying to smuggle a premise in the argument without anyone noticing.

Most people in the thread started arguing why it’s not offensive, while the onus was on the other side to prove that it was offensive. They had the burden of proof, and they inconspicuously shifted it.

If somebody starts a debate accusing you of racism, you already lost, especially if you try to defend yourself.

Sorry progressives, the word “master” is not “offensive terminology”. That’s what you have to prove. “What kind of project defends offensive terminology?” Is not an argument.

Adding a new configuration

This one is easy. There was no valid reason not to add a new configuration. In fact, people already had configurations that changed the default branch. Choice is good, this configuration was about making it easier to do what people were already doing.

The curious thing is that the only places in the thread where the configuration was brought up was as a diversion tactic called motte and bailey.

What they started with was a change of the default branch, a proposition that was hard to defend (bailey), and when opponents put enough pressure they retreated to the most defensible one (motte): “why are you against a configuration?”

No, nobody was against a new configuration, what people were against was changing the default configuration.

Should the name of the master branch be changed?

This was the crux of the matter, so it makes sense that this is where most of the time debating was spent. Except it wasn’t.

People immediately jumped to the next point, which is what is a good name for the default branch, but first it should be determined that changing the default is something desirable, which was never established.

You don’t just start discussing with your partner what color of apartment to choose, first your girlfriend (or whatever) has to agree to live together!

Virtually any decision has to be weighted in with pros and cons, and they never considered the cons, nor established any real pro.

Pro

If the word “master” is indeed offensive, then it would be something positive to change it. But this was never established to be the case, it was just assumed so. Some arguments were indeed presented, but they were never truly discussed.

One argument was that in the past (when slavery was a thing) masters were a bad thing, because they owned slaves, and the word still has that bad connotation.

That’s it. This is barely a argument, because not only is very tenuous relevant in the present moment, but it’s not actually true. Slavery was an institution, and masters simply played a role, they were not inherently good or bad, in fact some masters treated their slaves really well, to the point that many didn’t want to stop being slaves. Just like a boss plays a tyrannical role in principle, in practice many bosses are a delight.

Edit: A lot of people are are hanging on this point, and misunderstanding it. Slavery was bad. That doesn’t necessarily mean that all slave owners (like George Washington) were complete monsters, nor does it mean the word “master” had any negative connotation back then. It is an assumption we are making in the present, which, even if true; it’s still an assumption.

This is called presentism. It’s really hard to us to imagine the past because we didn’t live it. When we judge it we usually judge it wrong because we have a modern bias. How good or bad masters were really viewed by their subjects is a matter for debate, but not in a software project.

Supposing that “master” was really a bad word in times of slavery (something that hasn’t been established), with no other meaning (which we know it isn’t true) this has no bearing in the modern world, where the word is used in a variety of contexts:

  1. Recording (master record)
  2. An expert in a skill (a chess master)
  3. The process of becoming an expert (mastering German)
  4. An academic degree (Master of Economics)
  5. A largely useless thing (Master of Business Administration [MBA])
  6. Online tournament (Intel Extreme Masters [IEM])
  7. US Navy rank (Master-at-Arms [MA])
  8. Senior member of a university (Master of Trinity College)
  9. Official host of a ceremony (master of ceremonies [MC])
  10. Popular characters (Jedi Master Yoda)
  11. Golf tournaments (Masters Tournament [The Masters])
  12. A title in a popular game (Dungeon Master)
  13. An important order (Grand Master)
  14. Vague term (Zen master)
  15. Stephen Hawking (Master of the Universe)

All these are current uses of the word, not to mention the popular BDSM context, where having a master is not bad at all.

Yes, there is one meaning which has a negative connotation, but so does the word “shit”, and being Mexican, I don’t get offended when somebody says “Mexico is the shit”.

So no, there’s nothing inherently bad about the word “master” in the present. Like all words it depends on the context.

Even if we suppose that the word is “bad” (which is not), changing it does not solve the problem, it merely shuffles it around. This notion is called language creep (also concept creep). First there’s the n-word (which I don’t feel comfortable repeating, for obvious reasons), then there was another variation (which ends in ‘o’, I can’t repeat either), then there was plain “black”, but even that was offensive, so they invented the bullshit term African-American (even for people that are neither African, nor American, like British blacks). It never ends.

This is very well exemplified in the show Orange Is The New Black where a guard correct another guard for using the term “bitches”, since that term is derogatory towards women. The politically correct term now is “poochies”, he argues, and the proceeds to say: “these fucking poochies”.

Words are neither good or bad, is how you use them that make it so.

You can say “I love you bitches” in a positive way, and “these fucking women make me vomit” in a completely derogatory way.

George Carlin became famous in 1972 for simply stating seven words he was forbidden from using, and he did so in a completely positive way.

So no. Even if the word “master” was “bad”, that doesn’t mean it’s always bad.

But supposing it’s always bad, who are the victims of this language crime? Presumably it’s black people, possibly descended from slaves, who actually had masters. Do all black people find this word offensive? No.

I’m Mexican, do I get offended when somebody uses the word “beaner”? No. Being offended is a choice. Just like nobody can make you angry, it’s you the one that gets angry, nobody inflicts offense on other people, it’s the choice of the recipients. There’s people with all the reason in the world who don’t get offended, and people that have no reason, and yet they get offended. It’s all subjective.

Steve Hughes has a great bit explaining why nothing happens when you get offended. So what? Be offended. Being offended is part of living in a society. Every time you go out the door you risk being offended, and if you can’t deal with that, then don’t interact with other people. It’s that simple.

But fine, let’s say for the sake of argument that “master” is a bad word, even on modern times, and always, and the people that get offended by it have all the justification in the world (none of which is true). How many of these concerned offended users participated in the discussion?

Zero.

That’s right. Not one single person of African descent (or whatever term you want to use) complained.

What we got instead were complainers by proxy; people who get offended on behalf of other (possibly non-existent) people.

Gad Saad coined a term Collective Munchausen by proxy that explains the irrationality of modern times. He borrows from the established disorder called Munchausen Syndrome by Proxy:

Munchausen Syndrome is where a person feigns illness, or medical conditions, to garner empathy and sympathy. Munchausen Syndrome by proxy is when you take someone who is under your care (your biological child, your pet, your elderly parent) and then harm that person or entity that’s under your care so that you can then garner the empathy and sympathy by proxy.

So you see, Munchausen is when you feign illness to gain attention. Munchausen by proxy is when you feign the illness of somebody else to gain attention towards you. Collective Munchausen is when a group of people feign illness. And collective Munchausen by proxy is when a group of people feign the illness of another group of people.

If you check the mugshots of BLM activists arrested, most of them are actually white. Just like the people pushing for the rename (all white), they are being offended by proxy.

Black people did not ask for this (the master rename (but probably many don’t appreciate the destruction of their businesses in riots either)).

Another example is the huge backlash J. K. Rowling received for some supposedly transphobic remarks, but the people that complained were not transphobic, they were professional complainers that did so by proxy. What many people in the actual transgender community said–like Blair White–is that this was not a real issue.

So why on Earth would a group of people complain about an issue that doesn’t affect them directly, but according to them it affects another group of people? Well, we know it has nothing to do with the supposed target victim: black people, and everything to do with themselves: they want to win progressive points, and be “on the right side of history”.

It’s all about them.

The careful observer probably has already noticed this: there are no pros.

Cons

Let’s start with the obvious one: it’s a lot of work. This is the first thing proponents of the change noticed, but it wasn’t such a big issue since they themselves offered to do the work. However, I don’t think they gauged the magnitude of the task, since just changing the relevant line of code basically breaks all the tests.

The tests are done now, but all the documentation still needs to be updated. Not only the documentation of the project, but the online documentation too, and the Pro Git book, and plenty of documentation scattered around the web, etc. Sure, a lot of this doesn’t fall under the purview of Git developers, but it’s something that somebody has to do.

Then we have the people that are not subscribed to the mailing list and are completely unaware that this change is coming, and from one day to the next they update Git and they find out there’s no master branch when they create a new repository.

I call these the “silent majority”. The vast majority of Git users could not tell you the last Release Notes they read (probably because they haven’t read any). All they care about is that Git continues to work today as it did yesterday.

The silent majority doesn’t say anything when Git does what it’s supposed to do, but oh boy do they complain when it doesn’t.

This is precisely what happened in 2008 when Git 1.6.0 was released, and suddenly all the git-foo commands disappeared. Not only did end-users complained, but so did administrators in big companies, and distribution maintainers.

This is something any project committed to its user-base should try to avoid.

And this is a limited list, there’s a lot more than could go wrong, like scripts being broken, automated testing on other projects, and many many more.

So, on one side of the balance we have a ton of problems, and in other: zero benefits. That’s a though choice.

Best alternative name for the master branch

Since people didn’t really discuss the previous subject, and went straight to the choice of name, this is where they spent a lot of the time, but this is also the part where I paid less attention, since I don’t think it’s interesting.

Initially I thought “main” was a fine replacement for “master”. If you had to choose a new name, “main” makes more sense, since “master” has a lot of implications other than the most important branch.

But then I started to read the arguments about different names, and really think about it, and I changed my mind.

If you think in terms of a single repository, then “main” certainly makes sense; it’s just the principal branch. However, the point of Git is that it’s distributed, there’s always many repositories with multiple branches, and you can’t have multiple “main” branches.

In theory every repository is as important as another, but in practice that’s not what happens. Humans–like pretty much all social animals–organize themselves in hierarchies, and in hierarchies there’s always someone at the top. My repository is not as important as the one of Junio (the maintainer).

So what happens is that my master branch continuously keeps track of Junio’s master branch, and I’d venture to say the same happens for pretty much all developers.

The crucial thing is what happens at the start of the development; you clone a repository. If somebody made a clone of you, I doubt you would consider your clone just as important as you. No, you are the original, you are the reference, you are the master copy.

When you have a complex hierarchy where you have soldiers pushing towards sargents, and sargents pushing to lieutenants, and lieutenants to the benevolent dictator, you have many upstream branches, but only one master branch.

The master branch is the one that rules them all, and it was very likely created when the leader of the project typed “git init”.

When Stephen Hawking was coined “Master of the Universe”, no one thought; “oh, that means the whole Universe is his slave”, or “he must be a dick”. No, we understand that that means; he has a pretty good grasp of it, just like a master of chess.

The master branch is the one that keeps all other branches in check.

None of the other names come remotely close to this understanding, nor “main”, not “default”, not “trunk”.

Not to mention the true nature of actual software projects, where people submitting pull requests or patches are in fact at the mercy of the people with the power of accepting them or rejecting them. This is a dominance relationship.

Alan Fiske proposed the relational models theory (RMT), in which there’s only four types of relationships: communal, reciprocity, dominance, and market pricing. It is the reason why you don’t just grab food from your boss’ plate, while you might from others; it’s not that kind of a relationship.

Wishful thinking is not going to change the true nature of the relationships. Naming the branch “fuzzies” isn’t going to change the fact that most people are going to be subordinate to whomever decides what goes into that branch.

The name “master” fits like a glove.

Culture war

This was the other topic where a lot of time was spent on.

I don’t want to spend too much time on this topic myself–even though it’s the one I’m most familiar with–because I think it’s something in 2020 most people are faced with already in their own work, family, or even romantic relationships. So I’d venture to say most people are tired of it.

All I want to say is that in this war I see three clear factions. The progressives, who are in favor of ANTIFA, BLM, inclusive language, have he/him in bio, use terms like anti-racism, or intersectional feminism, and want to be “on the right side of history”. The anti-progressives, who are pretty much against the progressives in all shapes or forms, usually conservatives, but not necessarily so. But finally we have the vast majority of people who don’t care about these things.

The problem is that the progressives are trying to push society into really unhealthy directions, such as blasphemy laws, essentially destroying the most fundamental values of modern western society, like freedom of speech.

The vast majority of people remain silent, because they don’t want to deal with this obvious nonsense, but eventually they will have to speak up, because this dangerous ideologies are creeping up everywhere.

For more about the subject I can’t recommend enough the new book of Gad Saad: The Parasitic Mind: How Infectious Ideas Are Killing Common Sense.

It really is a parasitic mindset, and sensible people must put a stop to it.

The impact to users

I already touched on this on the cons of the name change, but what I didn’t address are the mitigation strategies that could be employed.

For any change there’s good and bad ways of going about it.

Even if the change from “master” to “main’ was good and desirable (which it isn’t), simply jumping to it in the next version: Git 2.30, is the absolute worst way of doing it.

And this is precisely what the current patch is advancing.

I already briefly explained what happened in 2008 with the v1.6.0 release, but what I find most interesting is that looking back at those threads many of the arguments of how not to do a big change, apply exactly in the same way.

Back then what most people complained about was not the change itself (from git-foo to “git foo”) (which they considered to be arbitrary), but mainly the manner in which the change was done.

The main thing is that there was no deprecation period, and no clear warning. This lesson was learned, and the jump to Git 2.0 was much smoother precisely because of the warings and period of adjustment, along with clear communication from the development team about what to expect.

This is not what is being done for the master branch rename.

I also find what I told Linus Torvalds very relevant:

What other projects do is make very visible when something is deprecated, like a big, annoying, unbearable warning. Next time you deprecated a command it might be a good idea to add the warning each time the command is used, and obsolete it later on.

Also, if it’s a big change like this git- stuff, then do a major version bump.

If you had marked 1.6 as 2.0, and added warnings when you deprecated the git-foo stuff then the users would have no excuse. It would have been obvious and this huge thread would have been avoided.

I doubt anyone listened to my suggestion, but they did this for 2.0, and it worked.

I like to refer to a panel Linus Torvalds participated in regarding the importance of users (educating Lennart Poettering). I consider this an explanation of the first principles of software: the main purpose of software is that it’s useful to users, and that it continues to be useful as it moves forward.

“Any time a program breaks the user experience, to me that is the

absolute worst failure that a software project can make.”

Linus Torvalds

Now it’s the same mistake of not warning the users of the upcoming change, except this time it’s much worse, since there’s absolutely no good reason for the change.

The Git project is simply another victim of the parasitic mindset that is infecting our culture. It’s being held hostage by a tiny amount of people pushing for a change nobody else wants, would benefit no one, would affect negatively everyone, and they want to do it in a way that maximizes the potential harm.

If I was a betting man, my money would be on the users complaining about this change when it hits them on the face with no previous warning.

Read More

RELATED ARTICLES

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Most Popular

Recent Comments