Hi, can you explain me what a Binaries-based installation is and how does it work?

Thanks

Heared first.

Is it installation of binary software packages from cd/dvd/net.

Compared to download of source and local compiling to binary software and installation.

1 members found this post helpful.

A great example by-contrast is the Gentoo distro, which is (or can be) completely source-code based: you compile everything.

Or, Linux From Scratch (LFS), which is exactly what it says.

When you "compile everything," of course you can tailor your software very exactly for your CPU-model and all other characteristics of your machine. But, it takes time. Sometimes hours. Therefore, it's far more common for someone else to have compiled the material, using "typical" options, and to then make available the compiled outputs ... the binaries ... which you then install on your machine.

I would, by the way, strongly encourage you to look into both Gentoo and LFS ... perhaps do LFS, in a virtual machine of course, strictly as the learning exercise. You'll learn a lot. You'll become a more effective user of those binary-only distros because you'll know a lot more about how everything actually works. And, you'll have a lot more appreciation for what those distro-suppliers have to deal with when they do what they do so well. The distros purposely insulate you from needing to know the gory details, but, to know those gory details is very informative. I think I learned more about Linux from that, than from any other source of information. ("Use the Source, Luke!")

1 members found this post helpful.

Most Linux installs are Binaries-based. So it isn't often described that way, just assumed. The exceptional case is source based.

In source based, you start with some previous Linux install and/or with a minimal binary based install and you compile all the binaries from source. I have read (but don't really believe) the idea that you get a noticeable improvement from the compiler knowing the exact CPU the binaries will be used on. The big difference is that you can get newer versions of some packages, if you want those.

In an ordinary distribution, all the compilation is done within the distribution team. They select combinations of source code versions, such as a version of an application and a version of some library the application depends on, and build, test and distribute those in aggregated versions. So the newer application version is available only with some delay and only with some newer library version. So you can get into a big dependency mess if you want to go bleeding edge on some applications while staying with the distribution on others.

When you build from source, you have much more flexibility in mixing bleeding edge from one source package with stable from another. Not everything will be compatible and not everything that is supposed to be compatible will actually work. You might be the first to try some combination. So it doesn't let you mix and match versions with total flexibilty, just with a lot more flexibility than in a normal (binary based) distribution.

1 members found this post helpful.