Rust-Written Redox OS 0.8 Released With i686 Support, Audio & Multi-Display Working

[Waethorn]

We are talking here about the Redox OS (it's in the title). Redox OS uses a microkernel. You keep mentioning Android for some reason that makes absolutely no sense or relevance. What's wrong with your head, I can't figure out.

EDIT: I just can't believe what an idiot this Waethorn is. He just made me check the title of the thread, because I was thinking that I might have been posting in the wrong thread.

I stated that hardware vendors like to write drivers for closed source operating systems using Android as an example. I could've used Windows as an example too, but I had mentioned ARM because that's the biggest "Linux"-supported hardware base right now (I was mentioning hardware vendor support for GPL vs BSD license permissability), just like how Mac is the biggest Unix hardware base right now, which is a bigger desktop *nix user base than desktop Linux. MacOS is the most successful commercial Unix product ever, and XNU is NOT a microkernel either, nor is it GPL. Your little statement about microkernels is a deflection, and means nothing. I could literally make a chart and show you your arguments versus mine with checkmarks to who proved their point, and you wouldn't have any.

[Waethorn]

The first problem is that you can't use Android vs main-tree Linux as an example because Redox OS (a microkernel) can alow drivers to stay under closed-source licenses, unlike the main-tree Linux kernel, where manufacturers are forced to open-source their drivers.

That statement EXACTLY backs up my point. Android has a different license than desktop Linux, which is what makes it popular among hardware manufacturers, which I will more-specifically name as the ARM SoC makers. Those same manufacturers, I might add, that have drivers for their chips for Android, but refuse to open source them for desktop Linux because they don't want to disclose their IP. GPL is not attractive to them because of that fact. Android, Windows, MacOS, iOS, all with permissive driver code licenses, are all more popular with hardware manufacturers. Your original statement to say that hardware manufacturers prefer open source licenses has been sufficiently debunked.

Android is ALSO, NOT a microkernel OS. You bringing up the point that Redox is, contains nothing to support your counter-claim to my statement.

The second problem is that you claim that Android is popular only because of licensing terms, but you are disregaring all the other possible factors.

You are writing complete nonsense.

And yet you provide no "other possible factors". Who's writing nonsense now?

[Waethorn]

Let's get back to the original point of contention:

manufacturers love to write drivers for microkernels.

Where's your proof of this?

[Waethorn]

I think you should do to some kind of medical check-up. Your gears are malfunctioning.



Nope, the original point of contention is in an earlier post.

Your gears are malfunctioning. Please do a checkup.

Despite your abusive insults, you have nothing to back up your claim that "manufacturers love to write drivers for microkernels". There is no statistic that you can point to that backs this up whatsoever.

[Developer12]

What do you mean? Are you saying that because all current microkernels don't have many drivers (or perhaps: in current microkernels drivers are typically open source), that it will stay like that in the future? Like, some crystall ball? Perhaps I'm misunderstanding what you wanted to say.

What you are saying here appears to me as mostly irrelevant, but perhaps there is some misunderstanding going on. Can somebody here shine a light on what Developer12 is saying?

Allow me to quote your own words back to you:

The simple answer: Microkernels allow for a fully closed-source drivers; so manufacturers love to write drivers for microkernels.

This is bullshit. For the reasons why, see below: (also found in my previous post)

1) There isn't a single manufacturer you can point to who has written a closed-source driver for a microkernel OS.
Meanwhile,
2) There are plenty of userspace drivers in FUSE on linux, and none of them are closed source.

Meaning your assertion is completely without basis in reality. Quite the opposite in fact.

And most importantly, why is that? Oh, right:
3) Open source software is strongly favored in the marketplace. No closed-source driver will ever make it into the redox mainline and be distributed to users, while any open source alternative will be the one to receive community support. Being a microkernel doesn't change that.

And finally:
4) A hypothetical microkernel driver must call through the userspace interface, meaning they can be traced for easier reverse engineering than on say, linux or BSD.

[Waethorn]

That claim of mine is irrelevant and inconsequential; all my arguments are valid without it.

I just added that (half of a) statement for fun. You can't prove it or disprove it, but it doesn't matter.

I just did with factual statistics that you don't possess. Hardware manufacturers prefer software that doesn't have an open source license, certainly not GPL, since it offers no IP protection. They also don't care about microkernels, since it has no bearing on solving the first point. Marketshare, wherein those devices are actually made and sold in bulk, are proof of that. Google came up with a superior license option that OEM's like. And BSD did the same for Apple. As did Microsoft for their hardware partners. Redox is, in your argument, an open source microkernel, with support for closed-source drivers. And yet there are no hardware manufacturers lining up to write drivers for it. It didn't work for Plan9 or L4 either. And ARM licensees are not putting in drivers en masse into the Linux kernel. This is all indisputable. Neither a microkernel design, nor GPL licensing solves any problem that the hardware manufacturer cares about, and so it's a pretty conclusive statement that this is the reason why they don't write drivers for it considering what they actually do write drivers for in real world metrics.

[Waethorn]

Microkernels allow for a fully closed-source drivers.

1) Irrelevant, because current microkernels are lacking either the in-kernel support or the proper ecosystem that would enable it.

This is a cyclical argument that makes zero sense and that also contradicts what you said in 3)

2) Irrelevant, because FUSE is not as good as a proper microkernel, so serious drivers can't use FUSE.

Opinion not grounded in reality. I could send this off to a bunch of FUSE developers that would have your head on a platter for making such a statement.

3) Irrelevant, because in a microkernel, drivers are not part of the kernel, so closed-source drivers can stay out of Redox mainline.

You're wrong on the first statement. This is not the definition of a microkernel. And you didn't even read the original statement for the second.

4) Irrelevant, because obscurity of a closed-source driver can be better accomplished by other means.

Absolutely wrong.

[Waethorn]

You can't prove or disprove the claim "manufacturers love to write drivers for microkernels.". First of all, it is not a mathematical nor logical statement.

To paraphrase Seinfeld: Who are these "manufacturers"?

Second, even if all the fairy-tales that you wrote are true, my claim is: Microkernels allow for closed-source drivers. Therefore, the licence of the kernel or the entire OS doesn't matter, because a driver in a microkernel is just another application. So you are wrong, and everything you have posted is just some irrelevant nonsense.

Ah so you'd like to add another point: it's "the license of the entire OS now"!

Show us hardware manufacturers that made userspace drivers open source. We'll wait....

[Waethorn]

EDIT: OK, I had overdone it.
Listen, Waethorn, there is a serious discussion going here. Anyone can make a mistake, and I'm sorry for calling you names. But, please, can you let me discuss this with Developer12, it is important, much more important than anyone's ego here.

If you don't like my opinions on the subject: too bad. You're not going to shut me up with your condescending tone.

[Waethorn]

Don't worry about that. Also, this what you just wrote shows something about you.

That I don't bow to name-calling, condescending bullies? Good.