Item — HN Reader

> If most people are not using a tool properly, it is not their fault; it is the tool's fault.

I would say that is a reasonable criticism of git ... but I've seen the same thing in svn, perforce, cvs, and rcs. Different variations of the same issue of people not caring about the version history.

Since it's been a problem since the dawn of version control, it is either something that is part of all version control being a tool's fault that has been carried with it since doing ci, or it is something that people aren't caring about.

I feel this is more akin to a lack of comments in code and poor style choices and blaming the text editor for not making it easier to comment code.

1 comments

> problem since the dawn of version control ... a tool's fault ... or it is something that people aren't caring about.

At the start of my career I ended up in a UI position. Old school usability on the back side of a 2 way mirror.

The tool has lots of shortcomings: images, documents that aren't text, working with parts of repositories... These aren't issues faced by the kernel (where emailing patches is the order of the day). And these shortcomings have lead to other tools emerging and being popular, like artifactory, journaling file systems, and various DAM's.

Technology on the whole keeps stacking turtles rather than going back to first principles and fixing core issues. Auth (DAP, LDAP, and every modern auth solution). Security (so many layers, tied back to auth). Containers and virtualization (as a means of installing software...). Versioning is just one among this number. We keep stacking turtles in the hope that another layer of abstraction will solve the problem, but we're just hiding it.

One of the few places where we (as an industry) have gone back and "ripped off the bandaid" is Systemd... It's a vast improvement but I would not call it user friendly.

Usability remains a red headed step child, its the last bastion of "wont fix: works for me" being an acceptable answer.