I have often found that one of the primary reasons that some analytics specialists and data engineers don’t “like” using git is that they don’t really internalise the underlying conceptual makeup of git and just try to learn the main “commands”.
For example, someone will learn that in order to share their work with remote they need to “git add -> git commit -> git push” even though there’s a tonne of stuff happening during those phases: you are staging your work with “add”, creating a commit node in the tree with “commit” and syncing your local HEAD / branch with the remote with “push”.
What inevitably always happens is that, at some point, somebody encounters an issue (eg a merge conflict) and then has no idea what is happening because they don’t really understand version control - they’ve merely attempted to “memorise” git instead which makes it harder to resolve issues as and when they arise.
Great article!
I have often found that one of the primary reasons that some analytics specialists and data engineers don’t “like” using git is that they don’t really internalise the underlying conceptual makeup of git and just try to learn the main “commands”.
For example, someone will learn that in order to share their work with remote they need to “git add -> git commit -> git push” even though there’s a tonne of stuff happening during those phases: you are staging your work with “add”, creating a commit node in the tree with “commit” and syncing your local HEAD / branch with the remote with “push”.
What inevitably always happens is that, at some point, somebody encounters an issue (eg a merge conflict) and then has no idea what is happening because they don’t really understand version control - they’ve merely attempted to “memorise” git instead which makes it harder to resolve issues as and when they arise.
That's very insightful, Johnny. My experiece is very similar.