"Distribution packaging for #Linux desktop applications is unsustainable"
Did you mean systemd/Linux?
I tried to read that article, but the example it gave was installing the proprietary malware known as steam, so I stopped reading.
The fact is, the kernel, Linux's SYSCALL ABI for each architecture does not change once it is stabilized.
You can run a statically linked elf from 1998 that only uses Linux SYSCALLs and it'll work today.
But, generally binaries are dynamically linked for security and sanity reasons and so you can't just copy those between systems and expect them to work.
Even so, generally libraries like glibc are forwards compatible, but you can't expect the ABI to be the same after 20 years - only the API.
As a result, the real packaging solution is to compile from source and let the compiler and build scripts to the rest.
Gentoo-libre and Guix is an example of how to properly manage packaging.
Really, you only have issues with packaging when it comes to using proprietary malware, as you can't compile software you only have the binaries of, as a result, the real solution is to just not use proprietary malware.
๐ณ๐ฑ
> First, it will never be the year of the "Linux desktop", as Linux is only a kernel and doesn't operate on its own.
Did you mean systemd/Linux?
Cant say this makes sense. Obviously linux needs many things to run and doesnt run on its own. It needs systemd, gnu userland, a x86 cpu, a motherboard, electricity, etc.. we dont say "Today is the year of linux, systemd, x86, motherboard, electricity"... we just say linux.
why, because its the component we care about in this discussion. We are sayi g "is this the year desktops finally start using linux as its kernel".
You need systemd? Sysvinit...?
Thats what I said, you **dont** need systemd, there are several alternatives and sysvinit is just one of them, there are at least a dozen others,
@freemo @Suiseiseki @lupyuen
I was agreeing with you @freemo ๐
.
