EasyOS is designed from scratch to support containers. Any app can run in a container, in fact an entire desktop can run in a container. Container management is by a simple GUI, no messing around on the commandline. The container mechanism is named Easy Containers, and is designed from scratch (Docker, LXC, etc are not used). Easy Containers are extremely efficient, with almost no overhead — the base size of each container is only several KB.
This is just one of the details of EasyOS, an experimental Linux distribution that really does things differently.
Good to see that after relinquishing oversight of Puppy, Barry has moved on to another project. Also good to see some of the thinking behind “frugal” installation has been extended into containerisation. Whether or not I ever play with this, there should be some interesting and useful outcomes.
Feh. Another “Let’s be different for the sake of being different”. Didn’t people learn *anything* form the Windows Vista/Windows 8/Gnome 3 nonsense?
When is an improvement not an improvement? When it’s not an improvement. Case in point,the whole container nonsense. How many times do people have to reinvent the wheel?
Sigh.
But, we learn a lot by “reinventing”. I admire those do an experiment over and over when everyone tells them, there’s no point.
IMHO, learning is always a valid point.
Yup.
Well, depends on who “We” are. Humanity at large does not usually learn something the billionth time quick sort is implemented, but yest that CS 101 student does. The challenge is always trying to determine which ones will have a benefit to the largest community. Like, I’m shocked, SHOCKED, that Rust has become as good and widely used as it has. it wasn’t the first to say “Hey lets do something like C but not as implicitly dangerous” , but its proven itself to be a good design and very useful.
So yes its a new container technology, maybe its good? I don’t know, ask someone better at predicting the future than me.
There would be no beloved Windows 7 without Windows Vista. There would be no Windows 10 without Windows 8. And so on. To answer your question, yes, end users and developers both learned a lot from the items in your list. People keeping `reinventing the wheel`, if that’s even a good analogy, because there’s no such thing as a perfect wheel, and because needs/wants/expectations change over time. There no reason people *shouldn’t* try to improve, optimize, or improve existing `wheels`.
It all starts with an experiment, sometimes you make a new better something…. sometimes not, but either way a new bit of information materialises.
There is no need to fear change just because it exists.
Vista was more of a case of they upped the security by a lot, but annoyed users by prompting for even a screen resolution change. 8 was a blatant attempt to force everyone onto the Microsoft Store and kill 3rd party software installs, which luckily was so unsuccessful they had to give up. As for Linux, the Linux people need to study the term “social darwinism” alongside colonialism to see while they will always fail. Instead of getting behind winning ideas and coming together, they keep fighting amongst themselves and wasting time on developing ideas like this under the false belief that one of these ideas will be what finally makes them competitive to Windows, Mac, etc. Essentially at this point it is a very apt comparison to say desktop Linux is like the countries colonial powers regarded as barbarians, easy pickings as they are simply too busy fighting amongst themselves to ever become a serious threat to anyone else. They need to actually copy what works, not keep developing features no one wants, or features that are only solutions to problems that exist because they use Linux (AKA containers to solve problems of using package managers).
Isn’t Linux the physical manifestation of the EU in an OS?
I suspect this is a bit myopic, “no one” and “they” are sure signs that “their” problems aren’t “your” problems!
If I ignore OS politics I think there is a place for all these options, given diversity if nothing else brings some chance of evolutionary survival. In a future world likely to be driven by an internet of things lightweight secure OS seems worth at least some research.
Of course they aren’t my problems, I literally don’t have those problems as I stopped using [desktop] Linux (as does pretty much everyone else that tries it). If no one entails 99% of users and the other 1% tolerate the problems, it’s good enough to use for the argument.
Also it doesn’t matter what those iot things use, there are already plenty of secure and lightweight OSes to use, nothing will change because of them, just like Android hasn’t improved the situation of desktop Linux whatsoever. Just another thing to waste development time on regarding small edge cases while the bigger problems go unsolved as fixing them has more to do with policy rather than technical merit of what’s the most efficient way to do something.
If it’s a from-scratch container implementation, is it at least using kernel functionality like cgroups, etc? If not, what is the mechanism and why should anyone trust it?
Also, I’m confused about the distinction between SFS, PETget, and Easy Containers. If you have a aufs-based package system like SFS, why not fully commit to it like Haiku did with its union filesystem package manager? Why have a traditional package manager at all? And what SFS and PETget binaries run in Easy Containers? Is every binary run in a container, or just some based on how the user configures it?
Digging a little deeper…
> The pup_event scripts are to be found in /usr/local/pup_event. The main guy is ‘pup_event_frontend_d’, which is launched from /root/.xinitrc when X starts.
> It is a binary file written in BaCon (a BASIC to C translator), that detects hotplug events and maintains the desktop drive icons.
WAT. When faced with the problem of service management and system init, adopting a BASIC dialect and C transpiler is not the first solution that comes to mind. Looking at an example (http://bkhome.org/docs/pup_event/pup_event_ipc.htm), I would think Lua would be a much better choice.