Sadly, Florian Mueller’s scream of outrage (IBM breaks the taboo and betrays its promise to the FOSS community is not an April Fool’s joke. IBM has done what it swore not to in 2005 – picked up the patent weapon and aimed it to block an open-source project.
I was thoroughly briefed on this about ten days ago by Jay Jurata, a lawyer working for CCIA (Computer and Communications Industry Association), which is even now bringing an antitrust action against IBM over the matter. Assisting him was Jay Maynard, an Armed & Dangerous regular who happens to be the lead of a project called Hercules.
Hercules is an open-source emulator for IBM mainframes. Words cannot easily describe my degree of bogglement the first time Jay brought it up an a Linux laptop in my presence and I saw the unmistakable arcane runes of a 360 boot sequence – and in an old-school band-printer-style font, too. Now, after 11 years during which IBM nodded when its own employees used and contributed to Hercules, Big Blue has brought down the hammer.
The issue begins with z/OS, IBMs’s mainframe operating system. For Hercules to be support the trillions of dollars’ worth of mainframe applications out there, z/OS has to run on top of it so the applications can run on top of that. But IBM has long refused to officially license z/OS to run over non-IBM hardware or emulators. Recently, TurboHercules SAS tried to change that.
TurboHercules SAS is a small company founded by Roger Bowler, the same guy who launched Hercules itself in 1994. He passed Jay Maynard the project leadership in 1999 when it went open-source, and later founded a company aimed at selling z-series emulation for disaster backup and other uses. In the U.S. it’s aiming at government agencies with IBM mainframes that have been ignoring their backup mandates because, jeez, one mainframe is expensive enough, let alone a spare.
Bowler asked IBM to license z/OS for use over Hercules and got back a letter that waved around vague charges of “intellectual property” violations. The letter outright refused to license z/OS for anything other than true-Big-Blue hardware. CCIA responded on behalf of TurboHercules by filing an antitrust action alleging illegal product tying.
This is where matters stood when Jay & Jay first briefed me on it. At that time, I told Jay-the-Lawyer that if the matter remained a spat between two for-profit companies over proprietary software it would be difficult to get anyone in the open-source community very excited about it. But I also said that if (a) IBM were to take direct legal action against the open-source Hercules project, or (b) violate its patent-nonaggression pledge from 2005, that would be a different matter – time, then, to light the beacon fires and gather the clans.
We still haven’t seen (a); no summons on Jay Maynard’s doorstep, yet. But a few days later we got case (b), big time. IBM sent TurboHercules a letter alleging that Hercules commits over 160 separate patent violations. Crucially, two of those were patents that IBM explicitly promised never to raise against open source.
I told them I’d blog that, for sure, but didn’t want to do it on or too near April Fool’s Day. Which is why Florian Muller broke the story rather than me. Florian has his facts right, but his interpretation is…perhaps a bit overheated.
It’s not justified to conclude from this attack, as Florian does, that IBM has never been sincere in its alliance with the open-source community. IBM has ploughed some big bills in marketing money and development time into the success of Linux. I am in a position to know for certain that IBM’s legal posture in the SCO case was tuned to help us out rather than defending solely their own interests on the narrowest possible grounds.
On the other hand…I was invited to speak at an IBM internal technical conference back in 1998, when they were formulating their pro-open-source strategy. They’d been reading my open-source papers, and the key planners behind the new, pro-Linux strategy wanted their peers to hear the word straight from the source. And the first thing I said into that microphone was that it felt strange for me to be there, because I had been a hacker long enough to remember when IBM (not Microsoft) was the Great Enemy.
That got a laugh, then…but those bad old days could be upon us again. Because for all IBM’s intentions to reposition itself as the world’s biggest consulting and systems-integration house, mainframes still generate 25% of its revenues and 50% of its profits. It could be IBM is reverting to type…or it could be that something like the rear brain of a Stegosaurus is defending its flanks by reflex before the official brain up font has had time to catch up with what’s going on.
IBM has reached a critical juncture. For the last decade the company has has had it both ways – allied with the open-source community to grow its forward-looking services business, and creamed big profits off the long-since-paid-for z-Series technology. Now these two strategies are in conflict. Whichever way this ends, IBM will probably only get to keep one of them.
I’m still not greatly concerned with the purely anti-trust aspects of this dispute. But our community must call IBM publicly to account on its violation of its own patent pledge for the same reasons we mobilized to help IBM defeat SCO in 2003, because we’ve got to keep their laws off our code. The kind of suppressive fire by blizzards of patents that IBM is deploying now is not just a single project’s concern, but a very serious threat to open-source development in general.
We need to find a way to push back. Because if there’s a first time that patent aggression works to lock us out of a whole area of software development, it won’t be the last. IBM needs to learn, as SCO and others have before, that that betraying our trust and going to war against us has consequences severe enough to make it a very, very bad idea.
I don’t have a dog in this hunt. However, I will say I find it strangely backward. I understand that ESR is not advocating any particular action here, aside from lighting the fires against IBM’s patent pledge, which I think is great. However, it strikes me as entirely backward to advocate an anti-trust suit, and not speak out against the true menace here which is the whole freaking patent system. The patent system is the problem here, anti-trust law is an abomination.
As the saying goes, as you all dine with the devil, may I suggest you use a long spoon.
I chuckle at the naivety of anyone surprised at this – in the end, businesses exist to make money, and will only support open source software (or anything else, really) to the extent that it assists that purpose. When you get in the way of them making money, there goes the support. esr, you said, “Now these two strategies are in conflict. Whichever way this ends, IBM will probably only get to keep one of them.” I’m not sure where you get that idea from. Assume IBM keeps pushing this on the Hercules project. How does that stop them using open source software in their business services division? It doesn’t. All it does is upset people who are in the open source movement. But you can’t prevent them from using it.
From the linked article, “This proves that IBM’s love for free and open source software ends where its business interests begin.” Of course it does. And there’s absolutely nothing wrong with that – that’s the way business and markets work.
>How does that stop them using open source software in their business services division? It doesn’t
Of course not. But if they earn the hostility of a lot of techies affiliated with the open-source community, they’re going to stop getting help from inside the shops they’re trying to sell to and start getting resistance that hasn’t been there before. This is all about whether we can make the consequences of crossing us sufficiently painful.
>And there’s absolutely nothing wrong with that – that’s the way business and markets work.
Agreed. But IBM going back on the pledge it made in 2005 puts this a bit beyond normal business rough-and-tumble. There have to be consequences for that.
I don’t understand. IBM said that they wouldn’t sue, thus allowing people to enter into business arrangements knowing that there would be no lawsuit. At this point, it’s too late to change their minds, as they have caused people to rely on their promise not to sue. I don’t see how they can hope to prevail in a court of law — over those two patents. Over the hundred-and-some others? Well, sure, they only gave away a limited number of patents.
Not that I’m defending the patent system, or indeed ANY form of purely mental process. If we don’t have freedom of thought, then we have no freedoms at all.
> This is all about whether we can make the consequences of crossing us sufficiently painful.
I suspect you can’t, because there aren’t enough of you. Open source advocates are like the Libertarian Party, in a way. There’s much to be admired there, in portions of it, but in the end, not enough to get anything like this done. As for the two patents they’d sworn, I’d expect an amended demand letter that only eliminates the two they’d pledged, but this, of course, won’t change any of the outrage. And nor should it.
Big business and open source software are always going to be at best a very uneasy alliance, which is only understandable.
They may not have much of a (legal) choice in this. IBM can’t simply let open-source organizations use their patents free of charge and then try to enforce the same patents against for-profit companies. They simply can’t discriminate that way. Years ago, IBM’s PC business got gutted by the clone makers; I’m sure that they still have some oldsters around that remember that.
Apple has always tried to keep a tight rein on software developers, and fought like hell against hardware copiers. I never liked that, always preferring the idea of open platforms, but I realize that if Apple had gone that route, it would have been the road to bankruptcy.
I had originally thought that the pledge was against Linux development, and some speak of it as tied to Linux and GPL. I see from the document available on-line that it is Open Source in the OSI sense and involving licenses deemed by OSD to be open-source licenses as of a particular date now in the past.
I wonder if someone at IBM considers the antitrust action to be a tripwire: an attack on IBM IP (z/OS) that nullifies the non-assert as any protection of the entity that is instigating the antitrust challenge (the way the Eniac patent was attacked by Honeywell, for example).
I don’t understand the different jurisdictions and the nature of the action that TurboHurcules took. Just wondering if that is the basis for IBM abandoning the non-assert. (Or they could remove those few patents from the list before they go farther.) Just speculating.
>IBM can’t simply let open-source organizations use their patents free of charge and then try to enforce the same patents against for-profit companies. They simply can’t discriminate that way.
Actually, yes, they can. The law allows this.
# LS Says:
> They may not have much of a (legal) choice in this. IBM can’t
> simply let open-source organizations use their patents free of
> charge and then try to enforce the same patents against for-
> profit companies.
Actually, you are completely wrong about this. The can do exactly this if they want to. That is like saying you can’t let the kid walking home from school walk across your yard, or else anyone can walk across your yard. It is your property you can do whatever you want with it. (That is, of course, based on the logic of “intellectual property” which I totally repudiate.)
You might be confusing it with trademark law which is quite different. Trademark law is all about the meaning of words. “Pepsi” means a particular soft drink. If I make a drink and call it Pepsi, it is essentially a fraud due to the pervasive use of the word to mean something else. However, to make sure it is pervasive then the trademark must be vigorously enforced. So IBM cannot allow Eric to start an open source company called IBM, but they can let him do whatever he wants with their other putative intellectual property.
> Years ago, IBM’s PC business got gutted by the clone makers;
I doubt that is true. The clone makers were one of the reasons that Apple, with a better product, had 10% market share, and Windows, with an inferior product had 90%. Everyone likes a second source. However, IBM might agree with your sentiment.
The problem with open source is that it is a pandora’s box. Once the contents are revealed you have no hope of ever hiding them again. I’ve been writing software for 20+ years and it’s been a long time (10+ Years or more) since I’ve showed up somewhere that they’ve written everything from scratch. Sure there will be spats about proprietary code from now and again, but this particular thing makes me want to build a black box implementation of z/OS just because I can.
It just never occurred to me that people would pay for something like this… there really isn’t a huge technical challenge here.
However, it strikes me as entirely backward to advocate an anti-trust suit, and not speak out against the true menace here
Huh? If IBM is using its copyright and patent monopolies in order to maintain a possiby illegal product tying arrangement, then suing them under anti-trust laws is eminently fair.
IBM’s policy of only allowing zOS to be run on actual big iron strikes me as fairly draconian. If the Hercules project becomes a target, it seems that it should be a priority to develop an open source migration platform for legacy zOS applications (running on mini- and microcomputers of course, not mainframes) under the broad interoperability provisions of copyright law,
This should address the kind of market abuse which IBM was historically prone to, while leaving unaffected the revenues from applications which actually need the high availability and throughput features of mainframes.
I’m sure IBM can be sued for breach of contract if there is an express contract with the OSS community.
Another thing is that there is such a thing as “promissory estoppel” and I’m sure the OSS community can use that against IBM.
@Jessica:
“That is like saying you can’t let the kid walking home from school walk across your yard, or else anyone can walk across your yard. It is your property you can do whatever you want with it.”
In the real-estate sense, this is actually flawed. If you let the kid cross all year, every year, while he’s in grade school, say, then shut him off because he’s in junior high school now, you may discover that you’ve granted him an easement: the right to use your property in a particularly limited way, that you can not revoke arbitrarily.
There’s a similar construct in contract and tort law (estoppel, I think?). I don’t know about patents, but I strongly suspect it does. With respect to trademark, it’s defend it or it dies.
What about the legal principle of laches? Hasn’t IBM, by not enforcing patents that it knew a 3rd party was infringing on, given up its ability to enforce them now, out of the blue (as it were)?
As the saying goes, as you all dine with the devil, may I suggest you use a long spoon.
Indeed.
I’d really like to know where IBM intends to go with this. Assuming for a moment that this was passed by somebody with a clue (at least technically) internally and not just an auto-response lawsuit filed by intern lawyers, I’d like a little clarification on the following:
My understanding of mainframe systems is that their main benefits revolve around certain hardware advantages – mainly really good I/O, multi-processing architecture which allows for hot-swapping CPUs in some cases. In short, either robustness or performance. To add one more business case, “enterprise” hardware also can come with a really long support contract (years to decades).
If you need either of the hardware advantages of a mainframe system, an emulator isn’t going to do it. You might be able to fake some of the robustness requirements with a VMWare setup which can do hot migration of VMs based on physical machine failures, but at that point you might as well run an app written for a more traditional OS (like Linux) and get your reliability through the VM directly. No amount of emulation is going to *improve* on native hardware for performance.
So, where would the really be useful:
* For development/test systems. You know the ones which allow you to do proof-of-concept so you can justify the multi-million dollar purchase of real equipment to the 3rd level management. This is why Oracle hands out scaled back licenses for its DB technology. Use our stuff, then get other people to pay for it once it works.
* For emergency backup systems. This is basically a stop-gap measure. If the original system wasn’t needed, they wouldn’t have bought one. If they need it, they’ll buy a replacement. This allows a company to go from “we can’t afford two” to “we can buy one and have our insurance pay for a replacement if the building burns down”. Not optimal for the 24/7 five nines, folks though. Basically it lets customers with smaller budgets afford the really cool stuff.
* For dealing with legacy stuff. These are the customers that enterprise dealers hate. They never upgrade. They never buy the high-margin stuff. They want support for equipment that was “end-of-life”d before anybody who works in the support center was hired (or born). Sure, you might have a 5x margin on replacement parts sold. But you risk more in terms of warranty service than anything else. You also end up giving away free upgrades: you need a replacement 100MB disk? We don’t have those many more – here’s a 100GB disk for half the cost. Please go away and be happy. There’s little real money here other than in support contracts. Having the customer move to an emulated system means you don’t have to support them any more. I know that most of the enterprise companies would rather these customers go away than have to support them, but keep supporting so that they can claim that they will offer support in 15 years or whatever.
So, IBM might be able to gain a few extra sales from some of that, but in exchange lose out on the others at the margin. This strikes me as worse than RIAA math.
Assuming that they are rational actors, I’m missing something here. What is it?
IBM’s patent pledge can be found at http://www.ibm.com/ibm/licensing/patents/pledgedpatents.pdf Florian Mueller ( http://fosspatents.blogspot.com/2010/04/ibm-breaks-taboo-and-betrays-its.html ) references http://www.scribd.com/doc/29469085/IBM-letter-dated-11-March-2010-to-TurboHercules-SAS as the letter proving IBM is reneging their patent pledge.
Should the two patents included in both lists be a basis for OSS outrage or worry? 158 patents not on the patent pledge were
asserted to Mr. Bowler. Certainly being apprised of those not estopped from prosecution by the pledge can have the same affect as the entire list.
You can question whether or not IBM owes an apology to the OSS community at large for the inadvertent inclusion of the two patents (US 5,613,086 and US 5,220,669). You’d think IBM might issue a public re-affirmation of the patent pledge following Mr. Mueller’s release of the letter. The issue could be how to do so without ceding the authority to try TurboHercules’ case in the court of public opinion.
You could contemplate that in a post In Re:Bilski world there might be some chance that patents could not be applied to a software emulator. In a contrary view US 5,613,086 Claim 1 (the independent claim) looks like it could be avoided in an emulator. The independent claims (1, 3 and 4) of US 5,220,669 appear to be avoidable at the cost of emulator architecture. Contemplating the remaining 158 patents you could speculate on whether or not implying the amount of work to avoid them isn’t IBM’s intent.
IBM appears to raise the stakes from antitrust by including patent license costs in the calculus of any costs associated with running z/OS on an emulator. Patent rights which would have been extinguished by the purchase of IBM or licensed hardware might otherwise affect the price. Think of it as a demonstration that all the cost of running system z are not reflected in the license cost of a copy of z/OS. In negotiating this might serve as indication the cost is out of reach, an encouragement to work around the patents akin to Samba or a prelude to throwing the other riff-raff out of the bar should ‘realistic’ license fees be forthcoming. The response has been instead of working to avoid the patents or amortize patent license cost to figuratively pound the table.
Other than the two patents TurboHercules doesn’t appear to have an issue on which to base their case in the court of public opinion. Antitrust complaints and legally granted monopolies appear to be genuinely antithetical otherwise.
History abounds with examples and counter-examples. For instance, Digital’s UNIBUS interface specs and protocols were generally available. Years later Digital came out with the BI bus, bound mummy-tight with licensing restrictions. 3P UNIBUS devices were available all over the place, and that fact [probably] contributed to Digital’s growth in that niche. In contrast, 3P devices for the BI were not exactly thick on the ground despite the backplane’s advantages.
Lots of people have opined that these four facts are actually causally related and that the syllogistic conclusions are obvious. My POV being what it is, I suspect there’s at least some truth to that. However, I’m pretty sure that it’s also too simplistic and that “there are insufficient data for a meaningful conclusion.” So, openness and hardware as business sense? Go figure. Are patents now shackles on the invisible hand?
I was an IBMer for more than a decade, and involved in its FOSS activities that entire time. Based on that experience, I think esr’s stegosaurus analogy is the most likely explanation for this. But now that it’s out in the open, there’s gonna be plenty of shitstorm to go around. And I bet a lot of boardroom walls inside IBM are going to have some smears on ’em, too.
RoUS, I’d be genuinely curious to hear your thoughts on the internal tug-of-war (single or plural) that’s likely to happen within IBM over this. If you don’t feel comfortable expanding on that here, catch up with me offline; I’m easy enough to find…
I think the most likely explanation for the inclusion of the 2 specific patents that they promised not to sue over is that whoever made the list of 150+ patents forgot to check them against that list.
kabdib, but open source had a license to use the patents. other parties … not so much.
Glenn, that differs from state to state, and if the easement is by permission (as opposed to inaction), then it may be taken away. If I don’t have no trespassing signs, you’re free to trespass (under certain conditions not worth listing here) on my land. If I then put them up, you’re not free. You’re thinking of the case where I have no trespassing signs and then I suddenly enforce them. You still don’t have a case unless you meet a bunch of other requirements, e.g. I cannot have ever told you not to trespass, you cannot be sneaking around, other people must know that you’re doing it. It’s really hard to prove. Basically, it only comes into play when somebody isn’t acting as if they own their land, and you are acting as if you do. Putting up no trespassing signs (and maintaining them) is a sign of ownership that I think most judges wouldn’t allow use to create an easement.
I’m not a lawyer even though it sounds like I know the law.
I strongly suspect that this is a case of the front brain needing to curtail the rear brain, or perhaps a poorly designed “breaching experiment” to see what _would_ happen. I strongly suspect the prior. This reeks (to me) of some nearly automatic process that was executed without oversight. Now its out, its come to proverbial blows and we’ll see what they do.
I can’t conceive how anyone in the top ranks of IBM would think that this is a good idea. This has to be a mistake, or an egregiously irresponsible test of the waters surrounding the issue. I could be wrong, I often am, but this just does not make sense.
I think Dio Gratia makes an excellent point Eric. Rather than heading down the ugly unethical route of anti trust perhaps you should rather work in the realm of In Re Bilski and Flextronics. Most likely all these zOS patents are right in the software patent realm. Why not try to do the world a favor and get them invalidated rather than making the world a worse place by asserting offensive anti trust law. Getting rid of patents is the work of a libertarian, asserting anti trust law is the work of a statist.
To continue my lame cliches, may I remind you that he who lies down with dogs, gets up with fleas.
>Rather than heading down the ugly unethical route of anti trust perhaps you should rather work in the realm of In Re Bilski and Flextronics.
Eh? In what sense am I heading down that route? I haven’t urged the community to support the anti-trust action, just to put pressure on IBM over its patent-nonaggression pledge.
I didn’t even get on board for the Microsoft anti-trust case – when one of Eliot Spitzer’s myrmidons called to recruit me I told him I was opposed to anti-trust law on free-market principle and declined. Said the same thing to Dan Gillmor in an interview a few months later.; I remember because he was quite struck by it given my history of warfare with Microsoft, and praised my integrity in a tone of no little surprise.
Dio Gratia, IBM has not proposed any ‘realistic’ license fees for running zOS on generic hardware. I don’t know anything about Roger Bowler’s letter to IBM, but it would of been stupid of him to ask for a free ride.
And yes, IBM could have gotten a lot of revenue from licensing zOS on even a ‘no support, you’re on your own’ basis. The most charitable interpretation for their stance is that (1) they want to avoid blame for zOS instances failing due to flaky hardware and software platforms (2) they don’t want the headache of coming up with a whole new licensing scheme for running zOS on unmetered HW (AIUI, mainframe SW is licensed on a CPU cycle or so basis). The least charitable is that they’re maintaining a lock-in position which leads companies to buy big-iron HW they have no real need of in order to run legacy apps. This is also why the CCIA’s antitrust complaint is relevant here.
(Jessica Boxer and other Royal Libertarians, you may not like anti-trust law, but asserting it is not statist if state monopoly is already involved. Copyright and patent law are state monopolies, so using anti-trust law to offset them is fair.)
Jay Maynard, you can get a glimpse of the tug-o-war inside IBM by reading this post from an IBMer.
>(Jessica Boxer and other Royal Libertarians, you may not like anti-trust law, but asserting it is not statist if state monopoly is already involved. Copyright and patent law are state monopolies, so using anti-trust law to offset them is fair.)
I’m going to ignore the ethical claim except to note that I disagree with it, because there’s a practical showstopper here. Doctrine derived from case law says, at least in the U.S., that you can’t answer a patent claim with an antitrust objection.
Eric, I understand your position, but this situation is quite different from the Microsoft antitrust case. In that case, Microsoft actually pointed to the existence of Linux/KDE/Gnome as proof that they didn’t really have a monopoly position in desktop operating systems. They even brought up the design of KDE’s kfm/khtml platform as evidence that integrating the browser in the operating system made technical sense.
It didn’t work back then because Linux was a marginal platform and KDE was little more than a curiosity–but it would probably work today. Launching a patent offensive against their open source competitors would have severely affected that defense.
Unless there’s compelling evidence that this was an action initiated with the explicit consent of executive management, my bet would be on the “reflex filing by some junior lawyer” theory.
This isn’t a case of some giant going out hunting for targets to sue. TurboHercules sued first.
Imagine you see two people on the street, apparently having a disagreement. Then one of them pulls out a gun and points it at the other one. Before deciding who’s in the right, it’s worth considering _which one is wearing a ski mask_. If someone tries to rob you, you’d be justified in drawing a gun and (implicitly or explicitly) threatening to shoot them in order to stop the robbery, even though you wouldn’t be justified in doing so to a random stranger.
This is not an action against the Open Source community. It’s a counter-suit against a proprietary software company which sued IBM. Equating it with patent trolling — without any evidence that legal action is planned against the open source community per se — is exactly the same logical error as when gun banners fail to distinguish between self-defense and aggression.
>This is not an action against the Open Source community.
Wrong. If Hercules can be rendered useless by patent lockdown, so can an unknown and probably large percentage of critical projects. And the first time anyone succeeds at this will unleash lawsuit hell on us.
That’s why any attempt to assert software patents to restrict the distribution or use of open source has the consequences of warfare and must be treated as such. Regardless of who the nominal target is and how justified the action might be in other respects.
But we don’t know that this _is_ an attempt to do that, is my point. It seems more likely that the intent is “make the antitrust lawsuit go away”…especially considering that IBM is certainly aware of the open source Hercules project, and is not suing its owner.
Is it a well-advised use of IBM’s legal rights? Probably not. But until and unless there’s some action targeting either an open source project or a company that didn’t _initiate_ the use of force, I’m disinclined to say it’s an immoral use of those rights.
IBM never agreed to refrain from defending itself vigorously against aggression…merely to not regard open source development and distribution as a form of aggression. Developing open source tools is not aggression, but hauling businesses into court is. Surely you, of ALL people, can see the distinction.
>Surely you, of ALL people, can see the distinction.
I sure can. Where I disagree with you is about whether this constitutes targeting an open-source project.
guest Says:
> Jessica Boxer and other Royal Libertarians,
If by “Royal Libertarians” you mean “people who encourage and favor freedom over state oppression”, I’ll be happy to wear the T Shirt.
> you may not like anti-trust law, but asserting it is not statist if
> state monopoly is already involved. Copyright and patent law
> are state monopolies, so using anti-trust law to offset them is fair.)
Are you arguing that the ends justifies the means?
ESR Wrote:
> Eh? In what sense am I heading down that route? I haven’t urged
> the community to support the anti-trust action, just to put pressure
> on IBM over its patent-nonaggression pledge.
You mentioned anti trust in your original, though, I certainly grant that you did not advocate it. I applaud you for the fact that you are willing to stand by your principles on this matter. I hope Jay and the others against whom this action is being taken will join you.
It seems to me that open source is primarily a matter of righteousness. I know that in much of your public advocacy you have made more of a utilitarian argument rather than a moral argument as RMS has done. However, I think in your heart you also advocate open source as righteous as well as useful. To advocate a righteous cause requires a degree of righteous behavior on the part of the advocates, regardless of the barbarian tactics against which you fight.
>I certainly grant that you did not advocate it. I applaud you for the fact that you are willing to stand by your principles on this matter. I hope Jay and the others against whom this action is being taken will join you.
Jay can respond for himself on that. As for the others (Bowler and TurboHercules SAS), I think they’re in any-means-to-the-end mode.
>However, I think in your heart you also advocate open source as righteous as well as useful.
Sure. But I learned long ago that idealism is often the worst enemy of the idealists’ own goals, so I acknowlege my ethical motivation without allowing it to screw up my tactics.
>To advocate a righteous cause requires a degree of righteous behavior on the part of the advocates, regardless of the barbarian tactics against which you fight.
True, but too much can be made of it. If I don’t win, being righteous afterwards is not enough consolation for me.
This is what kids these days call a “dick move”, to be certain, but I don’t think Dr. Jekyll is going to turn into Mr. Hyde completely this time. That is how they lost the PC business. Keeping the goodwill of the Linux community is important, as Linux plays a crucial part of IBM’s mainframe strategy. Like a lot of large companies (e.g., Sony) IBM can’t decide on the level of openness they want. They’ll learn what’s good for them, though.
Jay Jurata, the CCIA lawyer guy, is reading this thread and corrects me on a minor factual point.
He says TurboHercules filed the complaint in Europe on behalf of itself. CCIA’s role was limited to provided assistance to TurboHercules, similar to the assistance it provides other members.
Remember that there *was* a period where IBM, due mostly to the anti-trust suit ongoing, was forced to license their software to mainframe clone-makers, primarily Amdhal IIRC. IBM was finally able to cut Amdhal off during the 31-64 bit transition, and/or because they got the antitrust suit more-or-less dropping in 1982.
More recently there was another vendor, PSI, which was shipping an integrated z-Series emulation system on Itanium (there may have been some old Amhdal work in that) – they sued IBM and vice-versa and IBM acquired them in 2007; needless to say, their products were never seen again.
Garrett:
Your assumptions are wrong. The _original_ advantage of the big iron was I/O throughput, but that advantage disappeared almost 20 years ago. For years IBM has been selling the AS400 platform as their premier mini computer (small big iron) because it would do most of what the big machines do without taking up an entire room to do it. Since the early 90’s the AS400’s advantage over other platforms (e.g. the RISC-based systems commonly used for AIX and other Unix varients) has been a marketing rather than engineering advantage. The RISC systems when they came out in the late 80s (we had a demo unit before they were released to market at my university) had a larger IO path, more memory and a faster processor than the 5 year old IBM mainframe the university used for it’s business and accounting applications, registration, library catalog, and all the Data Processing Dept. classes. And that was in one tower case. The only thing it was missing was a large array of redundant disks, something that was solvable for just a few 10’s of thousands of dollars. However, IBM would NOT allow the VM OS to be run on the system (rumor was that IBMs engineers had already recompiled the OS and environment for the RISC6000 platform, but Marketing killed the product because it would cost about 20% of an AS400 base system and be as functional and scalable as the big iron systems that cost 50x as much).
We had a “mini” computer from IBM that the school paid almost $500K for that had this magical bus architecture as its main selling point. It was a 64 bit variant of the PS/2 bus (MCA? I forget the name now).
Almost ALL the advantage of Main Frames is gone and has been for at least 10 years. Moore’s Law tells us that the computer I’m typing this on is multiple times more powerful than the Mainframes I used 20 years ago (in fact, my PDA is likely more powerful).
The reason the project is important is that there are a huge number of internally build applications that run everything from business accounting functions to air traffic control systems that is dependent on the IBM hardware this stuff emulates, and since it was largely designed for systems that were orders of magnitude smaller and slower than what you can get today in a modern server platform, emulation is perfectly acceptable strategy for porting to less expensive, easier support hardware. With modern redundancy technologies like VMWare (Ha! The mainframe set invented VMs, so I guess there’s nothing new under the sun :^) the reliability is significantly improved.
Now somebody needs to create a similar emulator for the Honeywell systems that perform process control in virtually every chemical plant, refinery, and power facility in the world. Many of those systems have run non-stop for decades, and a failure there is measured in millions of dollars per minute and in many cases lives.
I’m no fan of emulation (I’ve always hated Java because of the jvm) but at the end of the day, it has it’s place in the tool kit and this is a REALLY BIG WRENCH!!!
# esr Says:
> True, but too much can be made of it. If I don’t win,
> being righteous afterwards is not enough consolation for me.
Right, as long as your definition of “win” is not excessively narrow. Gollum was victorious by some definition of “win.”
Matt:
Except for one minor detail: the only proprietary software involved is z/OS and friends. TurboHercules SAS is a company that sells services, support, and systems with the same open-source Hercules at the core that you can download from the project website. TurboHercules SAS is not a proprietary software company.
Jessica:
Unlike Eric, I have no philosophical objection to antitrust laws, and regard IBM’s actions as obviously in furtherance of a monopoly – and illegally so, and therefore subject to getting hammered. Then again, I make no claim to being a libertarian, though I prefer libertarian solutions where practical.
Don:
The advantage of the mainframe has never been raw CPU, It’s been reliability, availability, and serviceability. Mainframe admins who just reboot systems to fix problems get fired; the platform is much better than that. I’ve long argued that a DP manager who runs his business on Hercules is nuts, not because the emulator itself isn’t good enough (it has been for at least 5 years, in my view), but because several of the other things that make mainframe computing so reliable are missing, such as hardware error-checking and a dedicated support organization that can be called on 24×7. That doen’t mean Hercules is useless; far from it. It does mean that there’s a niche it fills nicely, and other niches it doesn’t fit well at all.
Jay Maynard Says:
> IBM’s actions as obviously in furtherance of a monopoly
> – and illegally so, and therefore subject to getting hammered.
Yes, how typical of the statist situation this is. The problem is caused by the government — who have granted IBM a monopoly. Now the state rides in on its white horse to save us with anti-trust laws. So the solution to a government created problem is another government created problem. And we wonder why the government grows inexorably?
However, I wish you luck in your endeavors to counter this unjust action on the part of IBM. I would suggest, from a utilitarian point of view, that invalidating the patents would be a good way to go. Antitrust law is difficult, In Re Bilski is also difficult, but just about resolved, which is to say, someone else did the work. Software patents are looking very shaky at the moment, and frankly, big patent holders like IBM don’t want to set such a precedent. This might be enough leverage to get them to go away.
US 5,613,086 is IBM’s patent for virtual memory page locking, which is pretty much used in all modern operating systems and the Linux kernel is certainly no exception. It would also make sense that Hercules would need to make use of the techniques claimed in this patent, being that it must somehow emulate the virtual memory system. (Side question for Jay Maynard: how is the virtual memory emulation done, exactly? Does it just use system paging, or is there something more to it?) In any respect, AFAIK, they can’t sue an OSS project or an OSS project’s sponsor over it, because they agreed not to do that, but if they try (and maybe even get away with it), it’s not too much of a stretch to see them using it against othe open source systems software.
In any respect, the blizzard of patents strategy is nothing new for IBM’s Nazgul, it’s just business as usual. Anytime they want to stop something they don’t want (like z/OS running on something other than IBM hardware), the very first thing they are likely to do is to go through their patent arsenal and threaten the offending party with it. This is exactly the same modus operandi the Nazgul used when they blackened the skies of Lindon, Utah. The idea is to intimidate the hell out of their victim, perhaps with the result of an out-of-court settlement.
>This is exactly the same modus operandi the Nazgul used when they blackened the skies of Lindon, Utah. The idea is to intimidate the hell out of their victim, perhaps with the result of an out-of-court settlement.
Er. No, that’s not at all how it went.
Trust me on this. There’s a good reason I disappeared for two years.
> Are you arguing that the ends justifies the means?
The ends are the only thing that can justify the means. Different ends justify the use of different means.
However, it’s always important to consider what ends will actually be produced by the means. After all, the means, not “intent”, produces the ends.
Let me ask a frank question for all you OSS guys. It seems accepted beyond question in your clan that software patents are bad. If that is the case, why aren’t all patents bad? What’s so special about software that makes it different than, for example, mouse traps, aside from the obvious fact that that they have a more direct impact on you?
Jessice:
Uhm, no they didn’t. IBM’s monopoly was obtained through questionable sales tactics, then curtailed originally in a consent decree to which they were subject from 1959 into this century. Indeed, their latest monopolistic tactics were unleashed almost immediately after they were released from the consent decree. It’s almost as if they thumbed their nose at the Department of Justice, which agreed to let them out of the consent decree. If I were the DoJ, I’d be extremely unhappy about this.
The reason software patents are bad is twofold: First off, the patent system as it is now structured was designed for a time when innovation was done at a much slower pace and so a longer term was appropriate, and secondly, the system is broken because trivial, obvious things are being patented because the patent examiners don’t know very much about software. This disproportionately affects software because the pace of innovation in the software word is so rapid, and thus a patent can stifle work for far longer than it’s actually useful; it also affects software because far more people are writing code than laboring in other fields of endeavor where patented inventions come from.
>This is exactly the same modus operandi the Nazgul used when they blackened the skies of Lindon, Utah. The idea is to intimidate the hell out of their victim, perhaps with the result of an out-of-court settlement.
Can someone point out what this is referring to, for those of us like me who don’t know what this references? Thanks.
@Darrencardinal: Simple search in google will reveal what they are talking about.
Jess, I think the arguments often made are Jay’s argument that software patents are often granted for trivial ideas and that many of those ideas required negligible investment, while some hardware patents required very large investment. I agree with you that the case for patents generally is very weak, but I do think we have to think carefully about how pharmaceuticals will be affected by removing drug patents.
# Jay Maynard Says:
> Uhm, no they didn’t. IBM’s monopoly was obtained through
> questionable sales tactics,
Curious, I am not sure how a monopoly can be obtained through the mutual consent between two parties. How this can affect other parties is not entirely clear to me. However, evidently I am not as familiar with the history as you are. From what I have seen all monopolies come either from government grants, or suppliers who do a really good job. Everything else is just a moat to climb. (And innovators are particularly good at inventing boats, ladders, portable bridges, hot air balloons and scuba gear.)
> The reason software patents are bad is twofold:
(Summary: software evolves too quickly to work with the traditional patent system, and software patents are granted on trivia.)
All business moves faster, and there are many businesses that move just as fast as software. Electronics, on which there are abundant patents being one obvious example. Further, trivial patents are granted in all fields. For example, if you look at lock patents, you will find patents for adding a tiny nub to a key here, or some slightly different way of mounting a lock on a door there. Every time a bridge is designed, or a plastic dinosaur is molded, someone does something slightly innovative. If all these things got patented, and those patents were enforced, the economy would grind to a standstill. That doesn’t happen because the courts are so mind bogglingly inefficient. However, what it does do is means everyone is violating someones trivial patent somewhere, and so everyone is vulnerable to lawsuit. This makes lawyers rich, and everyone else much poorer. There is nothing special about software patents in that regard.
You complain about the triviality of software patents: can you offer any examples of software patents that are sufficiently non trivial that you would support the grant of that patent? Or are you claiming that every single software patent is trivial?
# Ajay Says:
> Jess, I think the arguments often made are Jay’s argument
> that software patents are often granted for trivial ideas and
> that many of those ideas required negligible investment, while
> some hardware patents required very large investment.
But isn’t the reverse true too?
> I agree with you that the case for patents generally is very weak,
> but I do think we have to think carefully about how pharmaceuticals
> will be affected by removing drug patents.
What makes you think you are so wise you can calculate that incalculable? I already cited the GAO study that said that patents were destructive to innovation in the drug industry. Personally, I advocate this simple thing: “Let a thousand flowers bloom; let freedom ring.”
Jess, it’s difficult to think of a case of a software patent that required the kind of large investment in research that Chester Carlson had to put into developing the Xerox copier over 15 years, for example. The irony is that he was a patent lawyer and came up with the idea because of all the paper copies he had to submit to the patent office. :) If you know of software patents that required the same level of investment, feel free to share. Regardless of such outliers, I agree with you that there’s no real case for hardware patents either, once you factor in all the costs of such a monopoly. As for pharmaceuticals, if it’s really incalculable, what makes you so wise that you think you can calculate it either? ;) You keep citing this GAO study and yet never linked to it. I took a couple minutes and googled it, only to find that it was a fairly superficial study, dealing mostly with drug price increases from 2000-8 (feel free to link to another study if that’s the wrong one). Of the 6 drugs used as case studies, only two were even patented, yet all had large price increases. I’m fairly certain drug patents can also be scotched, but considering the billion-dollar investments they require, dwarfing all other patents out there, I think it’ll take some thought to make sure the alternatives are viable.
Whoops, here’s the Chester Carlson link.
Jessica:
IBM’s monopoly in mainframe computing wasn’t always such. They started out back in the 1950s by using a host of monopolistic tactics, and got slapped down hard for it. The original consent decree that governed IBM’s actions for over 40 years was adopted because IBM refused to sell its equipment at all, and refused to allow anyone else to make equipment that could be connected to it or to maintain it (all enforced through clauses in the rental agreement). The consent decree forced them to make technical documentation on how their equipment worked available to others. This let a marketplace for compatible gear spring up. When System/360 was introduced in 1964, the market really took off, and that helped the architecture and equipment compatible with it on both hardware and software levels to thrive. That it was upgraded compatibly over the intervening decades helped, since it encouraged large investments in software for the system. (An application program written in 1968 for OS/360 will, more than likely, run with at most minimal alterations on z/OS today.)
Customers flocked to the IBM mainframe world. They weren’t forced to buy from IBM, since there were several vendors of compatible equipment. This state of affairs continued through the end of the century. Other vendors in the marketplace (known as PCMs, for “plug compatible manufacturers”) came and went, but there were always a few, the best known being Amdahl Corporation, founded by the guy who was the principal architect for System/360 while he was at IBM. IBM, under the terms of the consent decree, would license their program product software to anyone regardless of who made the system they were using, although they could and did require that users of PCM systems get support from their manufacturers if a software problem popped up there. Amdahl’s support services were generally regarded as even better than IBM’s.
In the late 1990s, things were still going along pretty much as they had been. Then, two things happened: first, IBM got themselves released from the consent decree in 2001, followed quickly by their introduction of the 64-bit z/Architecture. The two events together caused the two remaining PCM manufacturers to drop out of the market, citing immense costs (including costs to license IBM’s intellectual property) and dwindling markets. The result, now, is that a user who’s invested literally decades of work in IBM-compatible systems has a choice: spend immense sums on migrating away from them (often more than the company itself is worth), or buy from IBM. The customer is locked in not by a conscious decision, but by IBM’s actions in driving others out of the market. That’s what makes them a monopolist under antitrust law, and that’s why they’re legally prohibited from tying their software to their hardware…or at least that’s the argument TurboHercules, aided by the CCIA, is making (and an argument I agree with).
> >This is not an action against the Open Source community.
> Wrong. If Hercules can be rendered useless by patent lockdown, so can an unknown and probably large percentage of critical projects. And the first time anyone succeeds at this will unleash lawsuit hell on us.
I might be wrong, but I believe this is confusing intent with results.
The *result* of this action might be negative for the OSS community (where, by “might”, I mean “will”), but it’s probably not proper to call it “an action against” us unless you can prove scienter. And it’s not at all clear that anyone could; I agree with those who speculate that it’s simply knees (probably new ones) jerking within IBM.
# Jay Maynard Says:
> The customer is locked in not by a conscious decision,
> but by IBM’s actions in driving others out of the market.
Thanks for the detailed history Jay. I appreciate you educating me. Now, let me put it back to you in the terms that I understand.
IBM successfully created a huge market place for powerful reliable computers. However, they hamstrung themselves by using tight licensing agreements that restricted users’ choice. However, the other benefits of their product overcame this limitation. Although there were other competitors, the totality of the package, restrictions notwithstanding, was considered the best buy for the customer.
However, the government didn’t like this (no doubt due to big lobbying.) So they interfered in the market. They broke the agreements that IBM had put in place, which, as it turned out, since they were a bad idea in the first place, benefited IBM. IBM was not forced to pay the price for such restrictive behavior. However, the benefit to IBM redounded to destroying the competitive market, resulting in a lack of direct competitors, but hangers on and peripheral makers who depended on IBM for their lifeblood, like some ugly symbiotic organism.
This went on for a while, and then, unilaterally the government changed their mind. The whole marketplace based around this ugly market interference collapsed, and since the consent decree had already destroyed the competitive market, there was nobody to step in and fill the void. However, now the barriers have been removed, but because the mainframe market has shrunk so much and been replaced by the PC, it is hard for others to come along and justify reinvesting the massive sunk costs that IBM has, and has had repaid under the umbrella of the government consent decree.
Now, companies that use IBM are stuck. They assumed the government would interfere for ever, and did not properly ameliorate the risk that they would not. Consequently, they are stuck, and will be bled dry by IBM.
Here is an alternative scenario: The government does nothing. Customers eventually get sick of IBM bs, and so a competitive market emerges. This market emerges based on reverse engineering of the product, using various legal trickery to get round the nasty agreements. (Check out how the Apple developer agreement got published as an example of how to defeat this kind of thing.) The competitors make compatible and better products. IBM sues for patent and copyright infringement, but loose because of the new, so called “Boxer Amendment” to the US Constitution that eliminates patents and many copyrights from the US legal system. As the competitive market emerges, IBM is left with two choices: open up, or close their doors. Most likely they choose the former, and the world turns out to be a better place.
So there might only be one seller, but it is a monopoly because of gross interference in the normal market mechanisms, including both the consent decree (which sounds like it benefited IBM, even though it destroyed the competition), and patents and copyrights that prevent developing competing compatible products. So, I state my case again, most monopolies have the root in government interference or government grants of exclusive rights.
@Darrencardinal: http://en.wikipedia.org/wiki/SCO-Linux_controversies
Heh, for a couple seconds there, I was racking my brain trying to figure out when Barbara Boxer came out in favor of such a staunch anti-IP position as amending the constitution to get rid of patents and copyrights. :D
> Heh, for a couple seconds there, I was racking my brain trying to figure out when Barbara Boxer came out in favor of such a staunch anti-IP position as amending the constitution to get rid of patents and copyrights. :D
Me too. Jessica’s Law is also taken, but Jessica’s Amendment works.
Yours,
Tom
Jess, it’s difficult to think of a case of a software patent that required the kind of large investment in research that Chester Carlson had to put into developing the Xerox copier over 15 years, for example. The irony is that he was a patent lawyer and came up with the idea because of all the paper copies he had to submit to the patent office. :) If you know of software patents that required the same level of investment, feel free to share. Regardless of such outliers, I agree with you that there’s no real case for hardware patents either, once you factor in all the costs of such a monopoly. As for pharmaceuticals, if it’s really incalculable, what makes you so wise that you think you can calculate it either? ;) You keep citing this GAO study and yet never linked to it. I took a couple minutes and googled it, only to find that it was a fairly superficial study, dealing mostly with drug price increases from 2000-8 (feel free to link to another study if that’s the wrong one). Of the 6 drugs used as case studies, only two were even patented, yet all had large price increases. I’m fairly certain drug patents can also be scotched, but considering the billion-dollar investments they require, dwarfing all other patents out there, I think it’ll take some thought to make sure the alternatives are viable.