Ask HN: Is Git as the only SCM ecosystem healthy?
As far as healthy, it's actually surprising so many have pulled everyone into a travesty that is git.
Yes, it's fast, but it sucks at everything else ("suck" is a bit too hard, it's more that other tools are more pleasant to use).
Still,the question you pose is an interesting one. I am curious to hear from others who have ported their projects to use git? I.e. there are hundreds of programming languages, but git it is for SCM.
Is it because the investment in learning it is so high? Is it because there are multitude of hosting solutions where others are rare and few? I really wonder!
IMO it's a mistake to think of Git as a version control system. Git is a toolkit of parts that you can mix and match to build a version control system.
Using the toolkit directly is awkward and comes with a steep learning curve; many of the parts have sharp edges that can result in permanent loss of your work, others have deeply unintuitive purpose (e.g. the various ways to edit history).
Any productive scaled use of git requires wrapping it, whether with scripting or process or social norms or some combination of those. Only after this work is done do you end up with something it is meaningful to compare with another VCS.
Whether the ability to customize flow to your needs justifies the effort involved compared to other VCS is really something each organisation should decide for itself.
Is fossil the only living comparable?
https://www.fossil-scm.org/home/doc/trunk/www/fossil-v-git.w...
So I guess the masses are OK with this.
And what happened to the other DVCS contenders?
They largely became unmaintained. I'd still prefer to use bzr over git, and I do with Launchpad.net for code hosting, but for any work, git has "won".
I have fond memories of Hg/mercurial, and remember Bitbucket dropping support. Just seems strange in today's technolandscape to have anything as a single-sourced option. But clearly git did "win". Now wondering if that will generate a demand for those who don't like the monopoly to push something else. Guess the question is, what does a better version of "git" look like and does it have to be compatible to have a change to get off the ground?
I can't imagine anyone wanting a "better git". To be honest, I can't imagine anyone wanting git, but here we are.
I do not understand why so many people feel it's imperative to use one SCM tool for all projects, since learning the basics is very simple for all of them (except maybe for git itself).
I get it why commercial hosting platforms standardize on one tool (everyone will ask for git, so why double the work?), so I suppose the desire not to run your own code hosting is the main driver.