It's a really good question, and one I probably should answer more directly on the website and in the README.
As I wrote above, I have a lot of time for these other tools—in fact the creators of sprinkle (Marcus Crafter) and deprec (Mike Bailey) are both mates of mine here in Melbourne, Australia—and I don't want to just duplicate their work. I'm trying something different here. Deployment is definitely still important to me with babushka, but I'm trying to build a more general tool that makes it as fast and pleasant as possible to automate getting to some end goal.
And that's one of the other differences—babushka is goal-oriented, not task-oriented. Unless there's some specific incompatibility, you can install babushka on (for example) any Mac in any state with this
bash -c "`curl -L babushka.me/up`"
And then running this
babushka homebrew
Will get you all set up with homebrew. But what it does is achieve the end goal of having homebrew set up—which might mean doing nothing, or cloning a git repo, or possibly just doing a `git reset --hard` to drop a binary in place. It depends on the system. The upshot though, is that if you're working on some other problem and a new colleague needs homebrew, you can run one command and you're done, instead of getting sidetracked for 15 minutes.
In focusing on that, I sacrifice usability in other areas—for example, right now there's not really any such thing as an unattended run using babushka, because most complex dep trees require user input at runtime. It's still very much a hands-on process. And that's cool with me—there are already tools out there that solve those problems, and solve them well.
I'm looking at working on deployment with babushka more with @glenmaddern, @dgoodlad and some other Melbourne hackers soon. But I'm very conscious of all the existing tools out there, and I'm a strong believer in deferring to the right tool for the job whenever possible. So I'm looking to integrate with existing tools where it makes sense, and solving old problems in novel and awesome ways whenever I can. :)
As I wrote above, I have a lot of time for these other tools—in fact the creators of sprinkle (Marcus Crafter) and deprec (Mike Bailey) are both mates of mine here in Melbourne, Australia—and I don't want to just duplicate their work. I'm trying something different here. Deployment is definitely still important to me with babushka, but I'm trying to build a more general tool that makes it as fast and pleasant as possible to automate getting to some end goal.
And that's one of the other differences—babushka is goal-oriented, not task-oriented. Unless there's some specific incompatibility, you can install babushka on (for example) any Mac in any state with this
And then running this Will get you all set up with homebrew. But what it does is achieve the end goal of having homebrew set up—which might mean doing nothing, or cloning a git repo, or possibly just doing a `git reset --hard` to drop a binary in place. It depends on the system. The upshot though, is that if you're working on some other problem and a new colleague needs homebrew, you can run one command and you're done, instead of getting sidetracked for 15 minutes.In focusing on that, I sacrifice usability in other areas—for example, right now there's not really any such thing as an unattended run using babushka, because most complex dep trees require user input at runtime. It's still very much a hands-on process. And that's cool with me—there are already tools out there that solve those problems, and solve them well.
I'm looking at working on deployment with babushka more with @glenmaddern, @dgoodlad and some other Melbourne hackers soon. But I'm very conscious of all the existing tools out there, and I'm a strong believer in deferring to the right tool for the job whenever possible. So I'm looking to integrate with existing tools where it makes sense, and solving old problems in novel and awesome ways whenever I can. :)