Hacker News Reader

You're conflating different things. Intel is free to use MINIX which is exactly the freedom the creator of MINIX intended when selecting the license. You are not a party to that transaction.

That Intel uses this software for a purpose you disagree with is immaterial to this discussion. Had MINIX not been available under a permissive license, Intel would have found some other (but presumably more expensive) way to largely achieve the same thing. They did not come up with and implement the management engine because MINIX was BSD licensed.

> You're conflating different things. Intel is free to use MINIX which is exactly the freedom the creator of MINIX intended when selecting the license. You are not a party to that transaction.

We are not a first party in that transaction, but feel very much the consequences. Therefore it does involves others. Most actions have an influence on others and therefore there are responsibilities for the consequences.

If you sell lower quality concrete to a building company which builds a dam that break and kills many people, then you have blood on your hands and should be treated as such, even though you were not a first party in the dam construction.

> That Intel uses this software for a purpose you disagree with is immaterial to this discussion.

It is very much not immaterial. It _might_ be legally irrelevant, but it is not without importance.

> Had MINIX not been available under a permissive license, Intel would have found some other (but presumably more expensive) way to largely achieve the same thing. They did not come up with and implement the management engine because MINIX was BSD licensed.

I hope they need to spend a lot of money and energy and effort for taking others freedom away. The cheaper it is for them, the more likely they do it.

Your beef is with Intel and the ME (specifically, the inability to disable it), not with Dr. Tanenbaum, MINIX or the BSD license. There is no causal chain from these to the Intel ME.

It may feel good to have a target to direct anger at, but it's not the right target and it's counterproductive. Free software, including that under the BSD license, has made computing vastly safer.

[the BSD license] has made computing vastly safer.

You have to at least stop and wonder whether that's true, though.

Your point of view and their point of view are both valid.

Unfortunately, both of you can be right, and we all have to live with that ambiguity.

I don't believe he is conflating anything. The author of the letter puts forward the notion that using the BSD licence was a good decision because it maximised the probability of companies like Intel using it. The commenter is implying that there's nothing particularly good about Intel having the free ability to shove a closed source ME-11 into most of our computers, that using a different licence would have in fact stopped them doing so, and that, contrary to your point, may in fact have led to them releasing the source.

"not party to that transaction" and "immaterial to this discussion" have no place here, we're discussing an open letter and we're free to evaluate his position however we feel. If the author wanted to have a private conversation with Intel he was free to do so.

Right: non-copyleft free software makes it cheaper - and therefore more likely - for companies to screw people.

Tell that to Microsoft, Oracle, SUN, IBM, Novell and SCO.

Non-copyleft proprietary software is even better.

If you are worried about this un-vetted blob, it means you don't trust Intel; if you don't trust Intel, why does it matter that they have this blob running in 'god' mode? They ARE the CPU, so they aren't just god 'mode' they are god itself.

You are trusting the CPU to do what you ask it to do. If intel were to do something shady, they wouldn't need a un-vetted blob of code to do it, they could do it directy in the CPU itself.

I guess my point is that you have to trust the CPU manufacturer, whether they have this code or not.

You are partly correct, but I think you're overlooking some things. I agree that we realistically have to trust Intel that their CPUs will do what the code tells them to. However, in this case they're actually telling us that the ME chip has overriding control over the system, that we can't tell the chip what to do, and that they won't tell us exactly what it's doing at any time or allow inspect the code its running. That isn't the case for their CPUs. This also opens up the possibility of a third party either finding a vulnerability in the code that we don't have access too, or simply gaining access to Intel's code signing keys, and using it to attack our computers. That's possible if the code works as designed, whereas that does not apply to CPUs.

Yeah, I guess my comments was assuming the trust was of the type 'trust not to do something nefarious' rather than 'trust not to have a vulnerability'. You make a good point about the latter being a more serious concern in this case.

Maybe you've got it backwards? Maybe many people trusted Intel until they learned about the ME and how hard it is to switch it off, and the fact that if you switch it off you should use a special flag requested by the NSA that was previously unpublished.

I certainly trust Intel way less than before some of the more fishy details about the Intel ME got out.

Intel's grown so big that you can't just do that anymore. Who's to say there ever won't be one employee on his last day with less ethics than you.

I completely sympathise, however your argument is based on a horse that has already bolted. Un-audited binary blobs already exist and run on your computer. Now off course, there is still an argument that you may not want even more un-audited binary blobs running, but security is about the lowest common denominator. We have passed that point already. Resistance is futile!

The management engine is one of the hardest binary blobs to remove though. Other ones like wireless and graphics firmware can be avoided by picking the right hardware, but the CPU doesn't have as many competitive options.

At least the blob inside my wireless controller has no access to the data that the blob in the ethernet port sends to the NSA...

That little Quark core has access to pretty much everything.

BTW, I'd love to use it to collect very low level performance data on the main CPU.

This is why the GPL is so great IMO. It would force Intel to either release the source code or use something else.

It would also force Intel to give us a means to replace ME with something else which ultimately puts the user in control.

Only the GPLv3 would force Intel to do that. With v2 they can go the TiVo route and make the source available without giving out the means to make the hardware run your own binaries.

Which means they could’ve used Linux if they wanted to

Why can't Linux use gpl v2 or later? Because Linus doesn't want to?

Linux is version 2. Linus is adamant about not using version 3. IIRC, some contributions are v2 or later, but as long as a majority of the code is v2 only, the whole package is v2.

"IIRC, some contributions are v2 or later, but as long as a majority of the code is v2 only, the whole package is v2."

If even one component is v2 rather than "v2 or later", then (I'd assume) the whole project is v2-only, majority or no. It's also not clear if it's permissible to make contributions under "v2 or later" licensing terms, since those contributions are derivative works of Linux itself (which is v2-only).

This will almost certainly always be the case, since changing Linux's license to "GPLv2 or later" (let alone GPLv3) would require unanimous consent from every contributor (and seeing as how - IIRC - one or more said contributors are dead, that consent is thus impossible to achieve unanimously).

The only sensible way to do it is to locate all parts still licensed under GPLv2 and ask every copyright owner alive (or their estates) to relicense their contributions under GPLv2+ or rewrite them from scratch licensing them under GPLv2+.

And convincing Linus the rewritten versions are objectively better than the old ones.

Nope, because they wouldn't use it then. I am not sure why people think the GPL had so much leverage over larger companies that will just build something in house.

But large companies use quite a bit of GPL software. I'm checked my phone, my PVR, my car media system and proprietary GPS I have and all of them have an open source license document that includes some GPL code.

For a hardware company, the GPL just isn't that much of a competitive disadvantage, since most of the software is written to work with their hardware.

Please convince Nvidia of that. Their closed source drivers are a major hindrance in the Linux community. They keep them closed because they feel the driver source would reveal additional details about their hardware, giving up some of their advantage over competitors.

I'm not sure why they used Minix at all. Surely Intel has enough expertise to build a custom OS that does exactly what they need and nothing else? That would have been easier to control for them, and it would have prevented this Minix-related news, which adds only more attention to something they'd have preferred to keep quiet.

Good. At least it will cost them to develop in-house. Raising the expense of doing user-hostile things is one way to dissuade companies from doing them.

Which is exactly what Intel did, since Intel didn't use e.g. Linux. If there were no BSD based OS to use, my bet is that Intel would have continued to use a closed source OS.

Even if they used Linux, they could just go the TiVo route and not release the signing keys

> or use something else

Which they did and if there was no alternative, they would develop 100% of it in secret. I think this is the opposite of great.

Absolutely. This is my first thought, too. What a ridiculous idea, that we have somehow achieved maximal freedom by giving Intel the legal means to hide from me what I'm being forced to run against my will.

You are not the user in case, Intel is.

Intel is not the user. They are a vendor, a publisher.

By calling them "users" and saying "users have freedom" you are conflating things needlessly.

We are the users here. We severely lack freedom. (Thanks BSD!)

Using copyright laws to force a person to release the fruits of their labor is not a definition of freedom that is widely agreed upon.

You are confusing things here. It is not about releasing fruits of labor. Intel is screwing us over and it is about stopping them.

Its not as black and white. At my company, the company owns the machines, and we have been looking at using Intel's AMT internally. I don't think Intel is screwing us.

That is cool. Good for you. I don't want to steal anything from Intel. I just want a way to opt out of ME and its siblings.

You seem like a troll or a bittered person.

Depends on who the user is. In a company, the company owns the computers, not the users. I'm willing to bet that ME was designed for legitimate corporate deployment scenarios.

You're perfectly free to not use Intel processors, and this 'GPL vs the world' argument is orthogonal to the BSD license.