Hacker News Reader

Comment by logicalmind | Hacker News Reader

This is mentioned elsewhere, but the connotation of master in the context of source control seems to be more along the lines of master used in terms of artifact creation. For example, where you have a master copy of a film or album. This is quite different from the definition of master in master/slave terms. This is definition 27 on dictionary.com.

   of or relating to a master from which copies are made: master film; master matrix; master record; master tape.

https://www.dictionary.com/browse/master

View on HN

The master terminology in Git is very likely historically tied to the master/slave metaphor, originally from bitkeeper.

https://mail.gnome.org/archives/desktop-devel-list/2019-May/...

But a branch from master is not a slave by any definition. Whereas a branch is a copy of a master. To be clear, I'm not against them changing this. But it just seems odd to base it on the implication that they meant it as master/slave when it clearly isn't used that way by anyone.

I was responding to your very general claim that "the connotation of master in the context of source control..."

Here is the bitkeeper documentation[1] saying "we are then going to modify the file on both the master and slave repository and then merge the work."

Safe to say that's in the context of source control? If you meant to say "based on how git works..." then sure, that's a more accurate claim.

https://github.com/bitkeeper-scm/bitkeeper/blob/master/doc/H...

I agree that bitkeeper should change from master/slave if they so desire. I'm not aware of anyone using master/slave in terms of git or any other mainstream vcs. And github's specific use of master does not use slave. By changing it, they do seem to imply that they meant it in terms of master/slave though. Branches are not slaves to a master. In standard usage, a branch overwrites the master.

Reply on news.ycombinator.com