"Adjusting the name of the initial branch in new repositories"

[Emanuele Ciriachi]

Today I updated git, and I was greeted with the following new screen during installation:

I would like to mention that I find it completely and utterly unnecessary - even insulting of our intelligence - to suggest that "master" as a branch name is somehow "offensive" or "not inclusive". In fact, the particular choice of words used in the popup evokes a mindset that is influenced by a biased socio-cultural context.

I really think that git would be better off by leaving this sort of pandering at the door.

Emanuele Ciriachi

[Philip Oakley]

The original choice of the word "master" in the technical context was referring to that of the absolute master driving the disposable slave.

You have a choice to make.

[Andrew Ardill]

Hi Phillip,

On Fri, 11 Dec 2020 at 01:39, Philip Oakley <philip...@iee.email> wrote:

The original choice of the word "master" in the technical context was referring to that of the absolute master driving the disposable slave. 

In the context of Git, this is not true at all. You can read the history of the usage of ‘master’ in Git in this article I wrote for the git developer newsletter:

https://git.github.io/rev_news/2020/07/29/edition-65/

--

Regards,

Andrew Ardill

[Philip Oakley]

Hi Andrew,

Your rev news article shows the history only as far back as the usage from BitKeeper, while I was referring to the original Horological usage that I referenced in https://lore.kernel.org/git/4bbc8658-4dad-10ef...@iee.email/ (see  Eglash: ref 4, also available via https://www.researchgate.net/search.Search.html?type=publication&query=Broken%20Metaphor:%20The%20Master-Slave%20Analogy%20in%20Technical%20Literature).

It is that master pendulum - slave clock analogy that has been propagated down through the technical usage and literature, carrying forward  the 18th and 19th century misguided beliefs.

For a dev's view maybe https://dev.to/gsto/how-software-can-be-racist-and-what-you-can-do-to-stop-it-1c3a (one of the comments has the link to the Eglash paper)

[Andrew Ardill]

Hi Phillip,

Appreciate the references, I think I read them when they were posted to the list but useful to have to hand. 

The article (or at least the research behind it - it’s been a little while) had two main points. 

1. The usage of `master` in Git did not derive from BitKeeper, but was chosen by Petr Baudis.  

2. The intention when choosing it was for a “master recording”, not related to the “master and slave” meaning of the word. 

I’ll also note that BitKeeper’s usage of the term was not related to slavery either. 

I understand that there have been many usages of “master and slave” in technology, and understand the efforts to remove those usages. You seem to be drawing a link between those usages and the usage of the word `master` in Git. I believe no such link has been demonstrated, there is good evidence that they are not linked, and the modern usage is not related to slavery at all. 

It’s possible that the historical “master and slave” usage has tarnished all uses of the word ‘master’ in technology for all time. If so, so be it, but this is a very different argument to the one you seem to be making – that having a master branch is “carrying forward 18th and 19th century misguided beliefs” – and there are far better (and more correct!) arguments than these if you want to change the name. 

--

Regards,

Andrew Ardill

[Andrew Ardill]

Apologies the below didn’t seem to render correctly. Here it is again:

The article (or at least the research behind it - it’s been a little while) had two main points. 

1. The usage of `master` in Git did not derive from BitKeeper, but was chosen by Petr Baudis.  

2. The intention when choosing it was for a “master recording”, not related to the “master and slave” meaning of the word. 

I’ll also note that BitKeeper’s usage of the term was not related to slavery either. 

I understand that there have been many usages of “master and slave” in technology, and understand the efforts to remove those usages. You seem to be drawing a link between those usages and the usage of the word `master` in Git. I believe no such link has been demonstrated, there is good evidence that they are not linked, and the modern usage is not related to slavery at all. 

It’s possible that the historical “master and slave” usage has tarnished all uses of the word ‘master’ in technology for all time. If so, so be it, but this is a very different argument to the one you seem to be making – that having a master branch is “carrying forward 18th and 19th century misguided beliefs” – and there are far better (and more correct!) arguments than these if you want to change the name. 

--

Regards,

Andrew Ardill

[Viktor Paulsen]

I agree. There is nothing wrong with "master". I have a Master of Science degree from a university. So has people of any gender and race. I am proud of being a master. I know many women and people of other races who are master. Please keep the identity politics and any politics out of git. Please don't be 'woke' and offended on behalf of others, about some made-up nonsense, in a git context.

[Philip Oakley]

Hi Viktor,

I see nothing wrong with the many types of "master of" qualifications which indicate a level of mastery of a subject or technique (such as a stone-masonry).

The difficulty is when the wording, for someone, is used in an absolute sense (I know all of science, I can never be wrong, my word is law, ...). Often it was used in the sense of 'the master of a house hold, it's good and chattels', of which the chattels were the wife (inc property), and the goods included the slaves. So it is delicate ground to tread on.

A 'master'  is also used in the sense of being an exemplifying (inanimate) object, such as a painting, a pattern, a drawing, from which copies are made of a lesser quality.

In terms of the Git data model, we already have perfect duplication, so the particular hash in my repo is identical to that in your repo, and that in anyone else's repo of the same project. Thus there is no single exemplary object. The distributed nature of Git means that any copy will do, so any continued use is just inertia from old bad habits of historical version control methods that we are meant to be free from.

Think of it as a grandfather's axe problem. Father replaced the head, now the handle's cracked and needs replaced.

Philip

[Viktor Paulsen]

> The difficulty is when the wording, for someone, is used in an absolute sense 

No, this is not difficult. This is artificial oversensitivity, pushed forward by political activists.

> So it is delicate ground to tread on.

No, it is not.

> In terms of the Git data model, we already have perfect duplication, so the particular hash in my repo is

> identical to that in your repo

That is not an argument of anything.

> Thus there is no single exemplary object.

That is not an argument of anything.

You can have a clone of the same master branch is as many clones as you like. It works beautifylly.

> any continued use is just inertia from old bad habits

There are no bad habits here.

Please stop the introduction of woke politics and artificial oversensitivity where it does not belong.

Here is the next task for you: "I think that some people might be offended by the word: git". It means "a foolish or worthless person". A foolish person felt very offended. It is delicate ground. 

You can spend your time renaming git. (nonsense, of course, presented here for humoristic purpose).