4 ms·

With all the complaining and name-calling, I'm quite surprised that nobody has proposed forking the project and maintaining a tracking-free version. Certainly,

With all the complaining and name-calling, I'm quite surprised that nobody has proposed forking the project and maintaining a tracking-free version. Certainly, it would appear as if there's a demand for such a project.

Conan_Kudo 8y ago
The pressure to fork Docker went away over the last year with the development of podman[1] and buildah[2]. There's some interest in the idea of wrapping podman/buildah into something that can be consumed by Windows and Mac users in a similar fashion to how Docker is right now. But it'll take some time to pull that off. [1]: https://github.com/containers/libpod https://github.com/containers/libpod [2]: https://github.com/projectatomic/buildah https://github.com/projectatomic/buildah
estesp 8y ago
What project? I'm assuming most of this commentary is regarding the Docker for Mac and Docker for Windows products which are an assembly of a significant number of open source projects: OCI's runc, CNCF containerd, CNCF Kubernetes, CNI provider, the docker/cli project as well as docker compose and kubectl binaries built for the host OS, the Docker CE engine built and packaged in a VM run by LinuxKit, xhyve, DataKit, VPNKit (all open source projects). I guess you could fork all those, but since many of them are not even controlled by Docker I'm not sure what the purpose would be. What I hope is relevant from the long list of projects that I just mentioned is that a company has spent a significant # of engineering years assembling, packaging, and supporting that combination in a way that makes it dead simple to do container-based development on non-Linux systems; mostly focused on developer laptops. No one else has that capability. It is a wide open field if anyone else wants to spend that same effort and time assembling a popular and free product that makes all that work together seamlessly on a Mac or Windows system. I am not saying I don't have an opinion on whether it's good or bad to make people sign in to download this free product. That is a company's prerogative who controls that product, and market forces will determine whether people will deal with such additions/changes. I of course would love to see direct downloads not impeded by such a change, but that's just my opinion. The silliness of HN is revealed when people start listing a bunch of other totally unrelated projects (cri-o, rkt, containerd) which don't provide any of the functionality of Docker for Mac or Docker for Windows. I say that as a huge proponent and maintainer of containerd. Again, if there is any other offering that makes that possible out of the box for Mac and Windows-based developers, then people are free to get behind that. To me, the only alternatives are to throw a VM together with Docker, Kubernetes, and whatever else you want and hack together the scripting and updating to make it work for you, and Docker nor anyone else are preventing or impeding anyone from doing just that.

Conan_Kudo 8y ago

The pressure to fork Docker went away over the last year with the development of podman[1] and buildah[2]. There's some interest in the idea of wrapping podman/buildah into something that can be consumed by Windows and Mac users in a similar fashion to how Docker is right now. But it'll take some time to pull that off. [1]: https://github.com/containers/libpod https://github.com/containers/libpod [2]: https://github.com/projectatomic/buildah https://github.com/projectatomic/buildah

estesp 8y ago

What project? I'm assuming most of this commentary is regarding the Docker for Mac and Docker for Windows products which are an assembly of a significant number of open source projects: OCI's runc, CNCF containerd, CNCF Kubernetes, CNI provider, the docker/cli project as well as docker compose and kubectl binaries built for the host OS, the Docker CE engine built and packaged in a VM run by LinuxKit, xhyve, DataKit, VPNKit (all open source projects). I guess you could fork all those, but since many of them are not even controlled by Docker I'm not sure what the purpose would be. What I hope is relevant from the long list of projects that I just mentioned is that a company has spent a significant # of engineering years assembling, packaging, and supporting that combination in a way that makes it dead simple to do container-based development on non-Linux systems; mostly focused on developer laptops. No one else has that capability. It is a wide open field if anyone else wants to spend that same effort and time assembling a popular and free product that makes all that work together seamlessly on a Mac or Windows system. I am not saying I don't have an opinion on whether it's good or bad to make people sign in to download this free product. That is a company's prerogative who controls that product, and market forces will determine whether people will deal with such additions/changes. I of course would love to see direct downloads not impeded by such a change, but that's just my opinion. The silliness of HN is revealed when people start listing a bunch of other totally unrelated projects (cri-o, rkt, containerd) which don't provide any of the functionality of Docker for Mac or Docker for Windows. I say that as a huge proponent and maintainer of containerd. Again, if there is any other offering that makes that possible out of the box for Mac and Windows-based developers, then people are free to get behind that. To me, the only alternatives are to throw a VM together with Docker, Kubernetes, and whatever else you want and hack together the scripting and updating to make it work for you, and Docker nor anyone else are preventing or impeding anyone from doing just that.