Hacker News Reader

Comment by codeulike | Hacker News Reader

Debate Question: would Linux have done so well without the wintel monopoly/monoculture on which it spread?

Depends on what we had instead. The fact that Wintel systems had unlocked bootloaders was absolutely critical. But it could equally have been started on any system with an MMU. Certainly it's been ported to a lot of systems.

Was it important to have a monoculture is what I'm wondering.

The fact that linux runs on almost anything would probably have helped it in a heterogeneous environment.

But is 'linux runs on anything' something that happened after it had achieved critical mass?

See sibling comment.

How long has that been true? It certainly started out as PC-only.

It was ported pretty early, I remember "beowulf clusters" of devices mentioned on Slashdot in the late 90s.

Edit: Checked wikipedia:

https://en.wikipedia.org/wiki/History_of_Linux#Chronology

1995: Linux is ported to the DEC Alpha and to the Sun SPARC. Over the following years it is ported to an ever-greater number of platforms.

> Was it important to have a monoculture is what I'm wondering

Obviously this is just speculation but I would say, no. And the reason I say that is that Linux primarily took off on the server side. Mobile was quite a lot later.

On the server side it wasn't a monoculture. Sun was a huge player. When we were making the case for Linux, around 2000, it was presented as a cost play against Sun/Solaris.

Of course, it's fair to say that part of the reason it was a cost play was because Intel kit was so cheap and that was a direct effect of the desktop environment. But I'd contend that it wasn't a reaction to the Wintel desktop monopoly.

Thats a really interesting view on it, thanks

You mean “IBM PC clone”. “Wintel” wasn't really a thing until Microsoft’s dominant consumer and business OSs were Windows, starting with Windows 95, by which point Linux had already done a lot of spreading.

Windows 3.0 and especially 3.1 were where it took over from DOS in the business world.

Windows 3.x required DOS 3+, it didn't replace DOS. NT and, on the consumer side 95, replaced DOS as the base OS.

That's not really an important distinction. 3.x was the first mainstream Windows.

Win95 still booted from a modified DOS.

True, but that DOS and Win95 weren't separate products.

It's because MS-DOS and to a lesser extent Windows 3 was so bad at isolating you from the hardware that all the hardware had to be identical. If MS-DOS had been more like, say, CP/M then we would have had loads of weird machines from different manufacturers isolated from the software by proprietary BIOS[1] blobs. This would have made it harder to provide a useful Linux that worked across multiple machines.

[1] I mean "BIOS" in the CP/M sense of a layer that completely abstracted the hardware to the BDOS.

Q-DOS from which MS-DOS is derived is basically a clone of CP/M. I've written a CP/M bios for fun to bootstrap my IMSAI. It's not really that high level. You got even more from IBM's BIOS but of course that was too slow for what people wanted to do with those machines - so everyone talked to the hardware directly.

In those days it didn't matter what you were doing - the hardware was too slow. It's not like today where you can trade off performance for reliability or cleanliness. People praise Wozniak for his genius of trading hardware for software, but the tight coupling made it really hard to make future iterations. People didn't value upwards compatibility yet.

Reply on news.ycombinator.com