MIPS Processor ISA To Be Open-Sourced In 2019

[schmidtbag]

Wise move. MIPS is on the verge of becoming irrelevant, so it'd be a greater loss for it to remain closed source and collect dust rather than heavily decrease its profitability while making it a viable alternative to ARM and RISC-V.

There is a lot of overlap, which is a real bad thing for MIPS, because ARM is hugely successful in comparison, and whatever is unappealing about ARM, RISC-V gets invested interest instead. However, there is a tiny place where MIPS doesn't have any overlap, which may heavily work in their favor:
Unlike ARM, they're open source, and unlike RISC-V, they're a well-established architecture with a decent amount of available software. For anyone looking for a fully open-source platform that is functional now, MIPS is really the only option.

[Dawn]

There is already POWER for high performance and RISC-V for embedded and low consumption devices. I think MIPS is poised to join OpenSPARC, OpenRISC and others in the also-ran category. Besides, I always felt that the MIPS ISA with its branch delay slots kind of sucks.

MIPS R6 uses compact branches, which have no delay slots. (It also has a bunch of other ISA niceties added, which really make the whole package a lot more appealing.)

[Markore]

Linux is not a fork of Minix. BSD license does not allow you to change the license of your product.

Unfortunately, (or happily, depending on thepoint of view), "BSD" license only stands there to say: "do whatever you want with this code". Including inclusion in proprietary, closed, products (Like it happend with using FreeBSD as basis for Sony's Playstatin 4 system software), and including re-releasing the same code under another license...

Why BSD-licensed software, continues to exist, thus unprotected with Copyleft clauses to force it to stay open, like GPL, LGPL, CDDL, Mozilla, Apache and others does?
Well because people contributing to it, like the way it is liberally licensed and including it's code routinely in proprietary products is not the issue for them..

With mentioning SPARC T1 and T2 hardware chip definitions (not T2+), being released under GPL long time ago,
they also depend on chipset for SPARC, thet is required for those chips to actually work in real-world hardware.
So that was the missing link, that didn't let SPARC , being released as Free software hardware definition, to flourish (Even Fujitsu is fabricating it's updated versions in it's and Oracle's servers).

[lucrus]

Linux is not a fork of Minix. BSD license does not allow you to change the license of your product.

Besides, I think Minix wasn't even BSD licensed at that time, since AST was selling it for money, albeit small amounts.

[starshipeleven]

Unfortunately, (or happily, depending on thepoint of view), "BSD" license only stands there to say: "do whatever you want with this code". Including inclusion in proprietary, closed, products (Like it happend with using FreeBSD as basis for Sony's Playstatin 4 system software), and including re-releasing the same code under another license...

Sorry what? You can't change the license of a work unless you are the copyright holder, and BSD license forces you to keep the original copyright owner in all files and make it clear also if you ship compiled code.

I mean, that's the only thing the BSD license enforces man, "don't delete the fucking copyright notice".

https://en.wikipedia.org/wiki/BSD_licenses

Why BSD-licensed software, continues to exist, thus unprotected with Copyleft clauses to force it to stay open, like GPL, LGPL, CDDL, Mozilla, Apache and others does?

Because people were stupid at the beginning, and it didn't backfire horribly by pure chance (not all projects matter enough to be a target). Does not make it a good idea. Next question.

[starshipeleven]

Besides, I think Minix wasn't even BSD licensed at that time, since AST was selling it for money, albeit small amounts.

Yeah, wikipedia says it became BSD way late, in 2000 AD, when Linux and FreeBSD were a thing already.

[Tomin]

Linux *is* a fork of Minix, or at least was in its early days. What you describe is exactly how it went. Yes, Linus prefers monolithic/hybrid kernels to microkernels so that was one of the first major changes he implemented.

I think it was more of an terminal emulator or something like that.

The important point though is that Linux doesn't contain a word of Minix source code.

[Tomin]

Yeah, wikipedia says it became BSD way late, in 2000 AD, when Linux and FreeBSD were a thing already.

Indeed. Minix had a propritary license that didn't allow redistribution back then. That was a major reason why Linux was developed in the first place.

[jacob]

Linux is not a fork of Minix. BSD license does not allow you to change the license of your product.

Historically it is, the BSD licence DOES allow you to change the licence of your own product (it's the GPL that doesn't), and Minix didn't use the BSD licence in the early 1990s.

[jacob]

I think it was more of an terminal emulator or something like that.

The important point though is that Linux doesn't contain a word of Minix source code.

Linux still contains a Minix FS driver. I don't know if that has been rewritten from scratch but for some reason I doubt anyone bothered. But otherwise you are right in that there is no Minix code left in Linux today, which is is why I said that Linux *started* as a fork of Minix (and then evolved totally separately). It's like Apache which started, IIRC, as a fork of the old NCSA web server. That's were its name comes from, by the way: "a patch - apache".