Will Git Be Around Forever? A List of Possible Successors by Hanno Embregts
Vložit
- čas přidán 5. 08. 2024
- Ten years ago, only Linux kernel committers and other early adopters used Git. Almost everyone else used Subversion. Ten years later, Git is the most popular product. Which makes me wonder: what will we use another ten years from now? And what features would YOU want from your version control software in 2032? No history rewrites? Faster? No merge conflicts ever?
In this talk I'll discuss a few post-Git products, including Fossil, Plastic and Pijul, and their support for the features we so dearly desire. I'll also try to predict which one will be 'the top dog' in 2032.
So attend this session if you're excited about the future of version control and if you want to have a shot at beating even (!) the early adopters. Now if it turns out I was right, remember that you heard it here first. 😀 - Věda a technologie
I must say I find Fossil almost magical and it seems to solve a lot of issues I have had with other SCMs. D. Richard Hipp is a true code hero in my book...not only for Fossil but also SQLite and AltHttpD.
Thanks for the comprehensive intro and demo to other version control systems!
THX for asking the difficult question: "What's coming for VCS?"
As I saw over my software career starting early 1980s:
Gen 0) No version control
Gen 1) Versioning by filename or folder
Gen 2) Centralized version control - one, central repo is the source of all truth
Gen 3) Distributed versions control - eventually consistent truth when we govern
Gen Next) ? Is storing patch instead of full result a generational change or a variant?
That is my core question: What kind of systemic change will replace distributed version control?
The past year has been fairly busy, with sapling being open sourced, jujutsu being an excellent alternative git client, and git-branchless becoming a very mature way to make git less painful
I think git is good enough to last for a very long time for most use cases. Most engineers are happy to keep something that works, everything changes fast in software so it makes sense to avoid switching to a new vcs if the gains are negligible(i.e. the time/effort for the switch probably won't be worth the benefits)
Good stuff. But SCCS existed on any standard Unix systems, much earlier. I had used it in early 90's. Not distributed of course, but quite workable for large C codebase.
I think it dated back to the 70s.
This guy is a slimmer version of Ryan Reynolds!
Pijul will be good when there's an email-compatible option for patch for folks not interested in creating accounts and when a simple self-hostable forge+web UI server options drop. Pijul also doesn’t support a rebase workflow to amend your commits before pushing.
The Darcs performance is a bit overblown. It’s not the fastest, but it’s improved overtime & folks don’t spend too much of their time in VCS tool specifically.
Let's just add Pijul integration to Source Hut!
And the fact it's so centered on patches themselves makes it infinitely more suitable for an email-based workflow than git is
@@mskiptr Pijul well never make it to SourceHut until it has a email workflow (even if just as a fallback). Darcs on the other hand would be nice as it’s ready to email & has the patch-based workflow. You can see in their separate request on the SourceHut mailing list where if the work were done, Darcs would be supported whereas Pijul would not.
Version control by USB drive? Fancy schmanzy new stuff. VC by 1,2 MB KB Floppy disk (AT 286).
I massively disagree with the lacking feature thing in git. It has all the features, be it built in or using third party software, which I think counts for git directly.
These new alternatives seem to try to do things different for different's sake also seen with the intentional diverging off the command names (you can alias git commands. I aliased "blame" to "congratulate". Inside joke with colleagues)
And people's issues with git are usually skill issues. Take a few hours to actually learn it, make a cheat sheet and you will rarely make mistakes or know easily how to fix it.
If you frequently mess up, then maybe you should check whether that hammer you thought you were using isn't actually a screwdriver.
Wrong! BitKeeper was not the only distributed version control system. There was monotone, and Darcs existed much earlier from around 2002-ish, but no one ever used that and it never went anywhere... But Pijul is taking a much more coherent way of managing your codebase.
I believe Arch was actually the first.
@@thadtheman3751 That sounds familiar... Wasn't it actually part of the Arch Linux project?
@@_general_error No. I think it was the first open source distribued SCM wriiten by Thomas Lord. The whole thing started when Linus needed help with the kernel organization. He went to Larry McVoy who wrote BitKeeper for him but made it proprietary unless used for on OSS project. He also amde it a condition that if you used it you could not contribute to a SCM project fort a year. larry Trigdell, the head of SAMBA/CIFS never used BItKeeper but used a client that allowed CVS users to accesws BitKeeper repos, wrote a better client. McVoy had a hissy fit and remove permission for OSS projects.
That spawned a bunch of SCMs. Arch (tla) was the first one. It ended when Linus wrote git.
what about support of those VCS in IDEs?
I had the same question. I think this might be a big issue for broader adoption - IDE and CI environments support.
I think i scripted „git all fetch background“ a dozent times
So if old is bad, this version control system published an hour ago must be excellent! The best! Until someone publishes a new one tomorrow!
What a closed mind
So if Netscape was bad, Internet Explorer must be excellent, the best! Until someone publishes Google Chrome. Then Firefox. Then Brave. Then Edge!
@@markmywords3817 I was being sarcastic, mate.
There are MANY version control systems better than Git, many of them that are unknown to most devs.
And they are better for technical and DX reasons. Wether they are newer or not is not the reason for their superiority.
Git is just popular but that doesn´t make it good.
I haven't watched the video.. But, this title is the kind of video, I am itrested in.
I mean, git shouldn't last forever.
It is good, but it's not that good
Git is the final version of version control. There won't be any other versions.