Hacker News new | ask | show | jobs
by pjmlp 3450 days ago
No, GNU/Linux needs less fragmentation.

It needs a full OS stack, from the kernel to the UI toolkit, that is guaranteed to be the same, across all distributions.

Where are the Mac OS X frameworks on GNU/Linux?

Nowhere, because each distribution might be using a total different way to play sounds, music, doing SIMD, process images, accessing databases, UI toolkit ....

3 comments

> that is guaranteed to be the same, across all distributions.

Then there wouldn't be "distributions", plural.

> Where are the Mac OS X frameworks on GNU/Linux?

Where are the macOS servers? Where is macOS running on a raspberry pi? Where is macOS running on your home router? On your watch? On your supercomputer? GNU/Linux does just fine.

macOS suits one use case, and one only (though it's very visible); don't make it sound like everything else sucks.

From Apple point of view those examples aren't interesting from business point of view.

If GNU/Linux wasn't free as beer and people had to pay for it, all those use cases would never have happened.

Hence why it is almost impossible to sell software to GNU/Linux users, and one has to get by writing books, selling consulting services or support.

Wait, macOS is free as in beer but all those use cases didn't happen with it, coincidence?
So where is the free macOS source code to build my own gizmo, profit and don't give any money back to those developers that helped me get rich?
Free beer means you get the beer for free, it doesn't mean you get the recipe.
This has actually always been one of the major benefits of FreeBSD (and other BSDs too).

I consider it an accident of history that Linux won out. We (devs, sysadmins, users) chose this fragmentation for some reason I still don't understand.

Choice, to me, is important.

I like that I don't _have_ to run a specific piece of software if I don't like the way it behaves/looks.

I can pick something else that gets the job done in a way I prefer

Don't kid yourself. If IBM had thrown its weight behind FreeBSD instead of Linux in 2000-2001, Linux would be the greener-grass "what if?" OS and FreeBSD would be on most of our phones. Individual users had nothing to do with it.
Even back then, Linux had more traction on embedded systems. FreeBSD was mostly desktop systems, though multiple architectures were supported.

The Sharp Zarus, Nokia stuff, and other efforts pioneered the way for Linux on phones.

Yes, that's actually the way I remember it around that time. People were just generally excited about hacking on Linux and the community as a whole had a sense of popularity about it that began in the 90s.

OTOH, I'm willing to accept the notion that, given all that, the big players still pushed the industry direction.

We all agree. How do you achieve that goal when device vendors have no interest in working with a distant third Linux in the OS ecosystem, much less your elegant but totally unproven PjmlpOS?
I don't, hence why I moved away from GNU/Linux as desktop OS.

For embedded use or being a POSIX clone on the server, it does just fine.