The decision to use the MIT License is not without its critics [LWN.net]

The decision to use the MIT License is not without its critics

Posted Jun 12, 2021 15:07 UTC (Sat) by BirAdam (guest, #132170)
In reply to: The decision to use the MIT License is not without its critics by Wol
Parent article: Rewriting the GNU Coreutils in Rust

Do you have any articles that fully explain your views on licensing and your choices for licensing your projects (esp your larger and more successful ones)?

to post comments

The decision to use the MIT License is not without its critics

Posted Jun 12, 2021 21:06 UTC (Sat) by Wol (subscriber, #4433) [Link] (15 responses)

Well, the project I keep on planning to write :-) will probably be licenced GPL2+.

But the EFFECT of that, seeing as it works with dynamic linking by the user at runtime, and subroutines are all independent modules, is that the GPL, in practice, is little different from Mozilla.

The only thing I would really do that's a bit different, is to stick a notice in the COPYING file that says "by distributing this software you agree that it is maths, a big number, software "per se", and as such that you agree it falls foul of the legal bars to being patentable subject matter".

So basically, anybody who distributes it is blocked from suing anyone else on patent grounds for using it.

Cheers,
Wol

The decision to use the MIT License is not without its critics

Posted Jun 12, 2021 23:30 UTC (Sat) by khim (subscriber, #9252) [Link] (10 responses)

<p>So you want to ensure that none of big players would ever touch it even with ten feet pole.

<p>I don't object to that desire but why not use GPLv3 then? It looks as if you would get all the problems of using GPLv3 without any advantages which is a bit of a strange choice.</p>

The decision to use the MIT License is not without its critics

Posted Jun 13, 2021 22:11 UTC (Sun) by Wol (subscriber, #4433) [Link] (9 responses)

From what I understand of all these patent clauses in the GPL, MPL, Apache etc, they all go on about granting licences to your patents. I can understand why big companies are scared of granting licences to all their patents.

But my version simply says "if you distribute this program, you can't sue the users with patents". It makes no statement as to whether the patents are valid, it doesn't grant any patent licences to anyone, it places (almost) no restrictions on who you can sue with them. It just says "you agree that this object is not patentable subject matter".

And like I said, chances are if I was going down the "patent licence" route, I'd choose the MPL not the GPL, because IN PRACTICE the two would be pretty indistinguishable.

Cheers,
Wol

The decision to use the MIT License is not without its critics

Posted Jun 13, 2021 22:42 UTC (Sun) by khim (subscriber, #9252) [Link] (8 responses)

> It makes no statement as to whether the patents are valid

Really? You basically say “software is a math, it couldn't be patented”, ask distributor to agree to that statement (WRT your program only, but what precisely makes it special?) and then expect that anyone who have any patents would just be happy to distribute it?

I don't really understand where such naivety may come from.

> It just says "you agree that this object is not patentable subject matter".

Which makes their whole patent portfolio much weaker (if not pointless). Because after they agreed to that they would have to explain what makes your software special and different from all other software. That's not an easy thing to do, especially if you need to convince not patent attorney but layman (which maybe needed in court).

And some think GPLv3 patent grant is too broad.

> And like I said, chances are if I was going down the "patent licence" route, I'd choose the MPL not the GPL, because IN PRACTICE the two would be pretty indistinguishable.

True. They both, at least, don't include any funny language which may be interpreted as a declaration to nullify other, unrelated, patents.

The decision to use the MIT License is not without its critics

Posted Jun 14, 2021 8:07 UTC (Mon) by Wol (subscriber, #4433) [Link] (7 responses)

> > It makes no statement as to whether the patents are valid

> Really? You basically say “software is a math, it couldn't be patented”, ask distributor to agree to that statement (WRT your program only, but what precisely makes it special?) and then expect that anyone who have any patents would just be happy to distribute it?

It's a special case :-) But basically, I just see it as stating the LEGAL reality that software IS unpatentable. Yes I know there are plenty of people tring to do an end-run round SCOTUS and the EPO treaty, but this would be my little contribution to the battle.

> I don't really understand where such naivety may come from.

Lawyers love arguing things. Judges rule as minimally as they can. Dragging this into a case about different software is quite likely to be thrown out as irrelevant, and if it's a case over this software, well, I doubt anyone would get themselves into a court case over it.

Maybe I just haven't been around big corporate lawyers enough ... :-)

> > And like I said, chances are if I was going down the "patent licence" route, I'd choose the MPL not the GPL, because IN PRACTICE the two would be pretty indistinguishable.

> True. They both, at least, don't include any funny language which may be interpreted as a declaration to nullify other, unrelated, patents.

And, in my case, as copyright licences, as I said they're pretty indistinguishable.

Cheers,
Wol

The decision to use the MIT License is not without its critics

Posted Jun 14, 2021 9:06 UTC (Mon) by khim (subscriber, #9252) [Link] (6 responses)

> Dragging this into a case about different software is quite likely to be thrown out as irrelevant.

How would that be irrelevant? If the company distributes your software and then tries to sue someone… how would that work? You can't simultaneously claim you believe software is just a math and then turn around and try to claim that it's patented. It's bad faith as minimum.

> Maybe I just haven't been around big corporate lawyers enough ... :-)

Or read enough news. When Oracle sued Google over Java copyright infringement was nine lines. And while, eventually, billions Oracle tried to get were never awarded… the case went to SCOTUS.

You forget (or misunderstand) what's the work of lawers actually is.

Their work is not to make sure you wouldn't be sued. That's impossible: is US, at least, anyone may sue anyone else for any reason.

Their work is not even to ensure you win. You are correct: there are many defenses which may ensure that court would (after few rounds of appeal) decide in your favor.

Works of lawers is to make court battles cost-effective! Otherwise why would they even get paid in the first place?

And your little stunt is bad as they come: some kind of unique (and very broad) patent statement they have to deal with in court. Sounds convincing to layman (at least to some laymans) and is not obvious why it should be ignored. Better to avoid that if at all possible…

> Lawyers love arguing things.

Only when they are paid to do that. In case of licenses they prefer to reuse analysis which was already done. They may not like GPLv3 but they know it. They discussed it with colleagues, they, most likely, already have a guideline and so on.

Your little stunt? They would ask for significant sum before they would be even ready to declare it safe or not safe! And in most companies this sum immediately becomes “price of that software” and often means that you can't use it. It becomes too expensive.

The decision to use the MIT License is not without its critics

Posted Jun 15, 2021 20:31 UTC (Tue) by Wol (subscriber, #4433) [Link] (5 responses)

> You forget (or misunderstand) what's the work of lawers actually is.

You forget I'm not an American.

> Their work is not to make sure you wouldn't be sued. That's impossible: is US, at least, anyone may sue anyone else for any reason.

Same here. Difference is, trying to sue your opponent into bankruptcy (popular with deep-pocketed Americans) is likely to end with you suing *yourself* into bankruptcy.

> Their work is not even to ensure you win. You are correct: there are many defenses which may ensure that court would (after few rounds of appeal) decide in your favor.

Or after just one round?

> Works of lawers is to make court battles cost-effective! Otherwise why would they even get paid in the first place?

Well, in America "cost effective" usually means "make the opponent lose worse", even if they are in the right. As I said, that's likely to prove seriously masochistic over here :-)

If I were a software distributor, and had no intention of using software patents offensively, there's a pretty easy defence over here.

If you sue me and claim that this means I agree software patents are invalid so I can't countersue, I promptly move for summary dismissal because you also agree that software patents are invalid, oh and can I also have you pay my court costs because you're suing me with patents you've just admitted are invalid therefore if it's not malicious it's negligent thank you very much.

Cheers,
Wol

The decision to use the MIT License is not without its critics

Posted Jun 15, 2021 21:25 UTC (Tue) by khim (subscriber, #9252) [Link] (4 responses)

> If you sue me and claim that this means I agree software patents are invalid so I can't countersue, I promptly move for summary dismissal because you also agree that software patents are invalid, oh and can I also have you pay my court costs because you're suing me with patents you've just admitted are invalid therefore if it's not malicious it's negligent thank you very much.

I think you have lost the forest for the trees. The distributor of your software agreed that software is a math and lost a chance to use software patents for a countersuit. His opponent haven't agreed to anything. Thus distributor couldn't use patents offensively, but is still under threat of patent lawsuits from others.

So yes, you have just gave the distributor another incentive not to touch your program.

Have you ever looked on any patent battles which were exposed on LWN? Why do you think patent lawsuits are usually settled instead of patents being invalidated? The answer is simple: because it's almost impossible to invalidate a patent! Once it's granted it's considered valid and you would need to, basically, go to the supreme court before you get invalidation will all rounds of appeals.

What happens instead is, usually, countersuit. And the best way to do it is to use your own patents. You don't claim that patent which was used to sue you is invalid. Instead you find some other patent which belongs to someone who you don't think would sue you and show it. Ideally an expired one. And claim that since that patent is also valid your software is not infringing.

Like: patent troll says that his innovation covers all the ways of drawing lines on a computer screen. You find a patent which explains how to draw lines using pixels and say to judge: “yes, plaintiff made great discovery and found a way to draw lines on a screen… that's certainly great and innovative achievement, but certain ways of doing that were known before — and here is patent that proves it”. And that's it.

Judge is happy because s/he doesn't need to pass through 10 layers of appeals, you are happy because you proved that you can draw lines on a screen… and even patent troll is happy because s/he can go and sue some other people because patent used for that is still valid (even if a bit damaged). And if he would agree to stop patent lawsuit and grant you license instead… then it's not even the least bit damaged! Without court decision it's not precedent!

This is the most cost-effective way of dealing with patents. And your “simple” clause makes it much, much harder (if not impossible) to use. Because now distributor have to somehow defend itself without bringing counterpatents to the table.

> Well, in America "cost effective" usually means "make the opponent lose worse", even if they are in the right. As I said, that's likely to prove seriously masochistic over here :-)

Maybe, but as long as you have to deal with US (and it's still one of the most lucrative markets) you have to deal with patents. Simply because you declared that software is not patentable doesn't mean patent trolls would agree.

Same here. Difference is, trying to sue your opponent into bankruptcy (popular with deep-pocketed Americans) is likely to end with you suing *yourself* into bankruptcy.

Oh, don't worry, that art would come to other parts of the world, surely. It's not even that hard: just sell your patent portfolio to dozen of pure trolls which can be bankrupted safely and see how many rounds of litigation the target firm would sustain.

You don't even need to actually coordinate lawsuits (and be accused of doing shady acts), just sell hundred patents to hundred patent trolls with list of companies which are infringing and make sure your target is there in top10.

The decision to use the MIT License is not without its critics

Posted Jun 16, 2021 8:53 UTC (Wed) by Wol (subscriber, #4433) [Link] (1 responses)

> > If you sue me and claim that this means I agree software patents are invalid so I can't countersue, I promptly move for summary dismissal because you also agree that software patents are invalid, oh and can I also have you pay my court costs because you're suing me with patents you've just admitted are invalid therefore if it's not malicious it's negligent thank you very much.

> I think you have lost the forest for the trees. The distributor of your software agreed that software is a math and lost a chance to use software patents for a countersuit. His opponent haven't agreed to anything. Thus distributor couldn't use patents offensively, but is still under threat of patent lawsuits from others.

You seem to be ordering me to give a damn.

Guess what, I DON'T.

> What happens instead is, usually, countersuit. And the best way to do it is to use your own patents. You don't claim that patent which was used to sue you is invalid. Instead you find some other patent which belongs to someone who you don't think would sue you and show it. Ideally an expired one. And claim that since that patent is also valid your software is not infringing.

Which is EXACTLY what I'd do. I'd let *them* claim my patent is invalid "because it's maths".

And because over here, the civil courts are courts of JUSTICE, I'd say to the Judge "please rule that software is maths, pretty please". Bang goes their case. And because we don't have the idiocy of doing ALL the discovery, THEN ALL the arguments, this would probably all be dealt with in summary judgement at the start of the case.

> Maybe, but as long as you have to deal with US (and it's still one of the most lucrative markets) you have to deal with patents. Simply because you declared that software is not patentable doesn't mean patent trolls would agree.

And if I don't give a damn? Plus, how long is it going to be the most lucrative? President Trump did a LOT of damage trade-wise to the US. China is growing rapidly in importance. Europe has been really spooked and will be seriously concerned about standing on its own two feet. India is a big and lucrative market.

America has been isolationist for most of its history. I think it's very soon going to find that Trump's mis-steps are going to haunt it and drive it back into isolation. On a personal level, the less I have to do with the US the better, and I think many Americans would be surprised how widespread that sentiment is. (Individual Americans, on the other hand, are nice. Including a lot of my relatives :-) But that's pretty normal for any country :-)

Cheers,
Wol

The decision to use the MIT License is not without its critics

Posted Jun 16, 2021 9:51 UTC (Wed) by khim (subscriber, #9252) [Link]

> Which is EXACTLY what I'd do. I'd let *them* claim my patent is invalid "because it's maths".

Why would they do that? They would be happy to add slander to your case. Because they still think that patents are valid and now you, too, agreed with that and that explicitly contradicts your previous steps (when you agreed to accept that code falls foul of the legal bars to being patentable subject matter). Court would need to decide what kind of outcome is actually valid (as usual: after years of appeals) but since harm is mounting (distribution software without license is big no-no) then can we have that travesty stopped?

Cease the distribution of that software and then we can continue. And that (preliminary injunction with the decision to stop distribution) may actually happen pretty quickly (few months).

> And because over here, the civil courts are courts of JUSTICE, I'd say to the Judge "please rule that software is maths, pretty please"

What law would give judge the right to say that? Courts may be “courts of justice” in the country where you live but they still can not invent rules on the fly, they have to follow law. That's the whole point of separation of power. Judges may think that software is math, it's their right, but as long as it's not part of some law they can not use that for court proceedings. The privilege of turning such beliefs into something you can use in court belongs to legislature branch, not judiciary.

Only high, final, courts in some countries have the right to “bridge the gap” and then, only in rare cases where one needs to “harmonize the law”, anyway.

> Bang goes their case.

Lol. I think you have never been in court. Not even as witness. “Bang goes their case” is almost never the case. And when you try to bring facts not explicitly mentioned in any law (as you repeatedly have tried to do in your mental exercises) appeals and repeated hearings may take years.

> And if I don't give a damn?

That's your right, of course. But as one of my friends likes to say “why would you pick such an extravagant way of self-mutilation?”

If you want to make sure others wouldn't use and/or distribute your software… you can just write so. No need to elaborate complex schemes which would lead to the same thing in the name of “fight against the software patents”.

> I think it's very soon going to find that Trump's mis-steps are going to haunt it and drive it back into isolation.

Possible. This would plunge the Europe into chaos, though, which may render question of software patents moot point because noone discusses software patents in Somali: they have more acute, more pressing, needs.

Do you really plan to release software with the goal of it only be usable after collapse of Western Civilization? That's a bit crazy.

> China is growing rapidly in importance. India is a big and lucrative market.

Indeed, China is growing and India may become more important with time, too. But they also fill patents like crazy. Including software ones. Do you really think they are doing this because they plan to continue to believe that software is a math and can not be patented?

Newcomers always try to somehow ignore patents. Incumbents always [try] to use them to protect itself. US was pretty negligent WRT copyrights and patents while it was growing in power. It only started to vigorously assert them when it couldn't grow.

If and when Europe would start failing (and without US backing it's inevitable) — it would double-down on software patents, don't worry. Same with China and India: if China would feel threatened — software patents would ineveitably become valid.

> Individual Americans, on the other hand, are nice.

Indeed. But individual Americans like individual Germans, individual Russians or, heck, even individual Iranians, never decide how laws of their country would be shaped. You need a political movement for that, not an obscure line in the license for the obscure software.

The decision to use the MIT License is not without its critics

Posted Jun 16, 2021 8:58 UTC (Wed) by Wol (subscriber, #4433) [Link] (1 responses)

> > Same here. Difference is, trying to sue your opponent into bankruptcy (popular with deep-pocketed Americans) is likely to end with you suing *yourself* into bankruptcy.

> Oh, don't worry, that art would come to other parts of the world, surely. It's not even that hard: just sell your patent portfolio to dozen of pure trolls which can be bankrupted safely and see how many rounds of litigation the target firm would sustain.

Unless it wouldn't ...

> You don't even need to actually coordinate lawsuits (and be accused of doing shady acts), just sell hundred patents to hundred patent trolls with list of companies which are infringing and make sure your target is there in top10.

And I wouldn't bother suing the patent troll for costs. At the slightest HINT of what you suggest, I'd go for the troll's directors and sue THEM. PERSONALLY.

Yup, it might prove expensive defending the first couple of lawsuits. But I doubt I'd have many trolls willing to put their personal wealth on the line ...

(I think "piercing the veil" with a patent troll company wouldn't be that hard.)

Cheers,
Wol

The decision to use the MIT License is not without its critics

Posted Jun 16, 2021 10:05 UTC (Wed) by khim (subscriber, #9252) [Link]

> And I wouldn't bother suing the patent troll for costs. At the slightest HINT of what you suggest, I'd go for the troll's directors and sue THEM. PERSONALLY.

So you would sue someone who's work is, quite literally, to be sued? Big deal.

The actual guys behind the whole thing would laugh all the way to bank.

> (I think "piercing the veil" with a patent troll company wouldn't be that hard.)

Have you ever tried to do this? Sigh. I wish you will read the The Little Golden Calf. Then you would know who Sitz-Chairman Funt is. Actual wikiquote may just be enough:

Sitz-Chairman (зиц-председатель, zits-predsedatel), a strawman chairman. The Horns and Hoofs was headed by Sitz-Chairman Funt. The title is a bilingual Russian-Yiddish pun. The Yiddish word "sitzen" means "to sit", which in Russian connotes "doing time". Also "Sitz" has legal meanings similar to the English "seat". The sole function of a Sitz-Chairman was to do prison time when (not "if"!) the time comes (Compare Sitz-redakteur, a person hired by a 19th century German newspaper for the same purpose).

Note the last phrase. It all was invented in good old Europe and, don't worry, it will be back sooner or later.

The decision to use the MIT License is not without its critics

Posted Jun 14, 2021 17:19 UTC (Mon) by ceplm (subscriber, #41334) [Link] (3 responses)

> Well, the project I keep on planning to write :-) will probably be licenced GPL2+.

Well, it doesn't make any sense to me. Why the plus? Either you are against GPL3, and then you want GPL2-only, or you don't care and then you go for GPL3. GPL2+ doesn't make much sense in this day and age.

The decision to use the MIT License is not without its critics

Posted Jun 15, 2021 19:00 UTC (Tue) by nix (subscriber, #2304) [Link] (2 responses)

Maybe you like v3 but don't want to stop v2-only projects from using you?

The decision to use the MIT License is not without its critics

Posted Jun 15, 2021 19:59 UTC (Tue) by Wol (subscriber, #4433) [Link] (1 responses)

Actually, I hate all this *patent* licence crap in what's supposed to be a *copyright* licence.

That's what I don't like about GPL3, although I understand why they put it in. Although I'm hoping that the Google/Oracle spat has helped clarify the fact that if it's copyrightable it's not patentable ... :-) At present I believe software is unique in that the lawyers would like you to believe it's both?

But GPL3 also has significant bugfixes to GPL2 - bugs that bite badly in an internet era. So GPL2+ suits me fine.

Cheers,
Wol

The decision to use the MIT License is not without its critics

Posted Jun 16, 2021 1:17 UTC (Wed) by mathstuf (subscriber, #69389) [Link]

> Although I'm hoping that the Google/Oracle spat has helped clarify the fact that if it's copyrightable it's not patentable

I don't think the case did that at all. Google won on fair use grounds which is reusable by precisely no one except as a benchmark to compare their reuse against. The patent claims were, IIRC, thrown out early, but kept it on the federal circuit because patent claims automatically go there.