There's a fundamental error in the InfoWorld analysis referenced above:
Oracle simply asks for absolutely standard remedies in this situation. There's nothing evil about it, and it cannot be reasonably interpreted as a strategy to destroy open source as a whole or anything like that.
I'm saying this even though I opposed Oracle's acquisition of Sun. I just want to point out that if a case like this goes to court, the plaintiff will always ask for those kinds of remedies. There's nothing unusual about it. In fact, asking for less would be unusual and would probably confuse the judges as to what Oracle actually wants.
Intellectual property rights are exclusive rights. That's the way the law has designed them -- it's not a matter of Oracle being evil. Those IPRs entitle a right holder to enforce exclusivity. That necessarily means to ask for an injunction, and under such circumstances as the ones of this case (with copyright in play), also the destruction of infringing material.
The way to prevent that scenario from materializing is a license agreement between Oracle and Google. So it's up to the two parties to sit down and negotiate, and I believe we as a community should now expect both of them to be constructive. The court can't impose a license agreement on the two of them. If the court has to rule, it will -- if Oracle is right -- have to enforce exclusivity. That's sort of binary, whereas a license agreement would offer much more flexibility.
It's regrettable that they couldn't work this out before the matter was taken to court. But it's not too late until there is a final court ruling.
There's some confusion out there about how Google would be in a better position from a patent point of view if it had used existing Java code under the GPL (OpenJDK, phoneME). The Android ecosystem as a whole would have had other benefits (such as making it much harder for the makers of Android-based phones to keep important parts of their source code closed) but it wouldn't really help as far as Oracle's patent infringement allegations are concerned.
The GPLv2 (under which OpenJDK and large parts of phoneME are available) does not contain an explicit patent grant. Only an implicit one. As a result, any fork (derived/modified version) is probably not covered.
The InfoWorld article that this Slashdot story refers to talks a lot about forking as a possible strategy -- especially toward the end, where MySQL is also mentioned. I was very much involved with the debate over whether Oracle should get to acquire MySQL (together with Monty, MySQL's original author/founder, I opposed the deal). In that context, it was also a subject of debate whether MySQL forks would be safe from Oracle patent threats in the future. Eben Moglen, who was basically part of Oracle's legal team and had botched the patent aspect of GPLv2 (thus tried to fix the problem with GPLv3), argued that GPLv2 would take care of those forks. However, the European Commission, which (unlike Moglen) is impartial and has vast legal resources, concluded that the implicit patent grant does not -- at least not reliably, but probably not at all -- protect forks.
If you're interested in more detail on the question of whether Google would be or would have been better off with GPLv2, here's a link to the related part of a blog posting of mine. It discusses that question and right thereafter (or you can go there directly) explains that my work related to Oracle's acquisition of MySQL was not an effort to change MySQL's license away from the GPL to something else. I have meanwhile published documents from the process that serve as conclusive evidence that I argued vehemently against -- not for -- a license change. Still, the GPLv2's limitation concerning patent claims against forks is a fact.
Indeed, Apache is covered, and so are lots of other things that go way beyond the Linux kernel. The problem is that the OIN has that completely arbitrary approach to determining what's on the list and what's not. I actually suggested four alternative ways to address that problem, but it seems they don't want to change anything. They want the OIN to continue to be totally intransparent (which in terms of litigation would be understandable, but the definition of the scope could be a transparent process). And they want it to simply benefit the six companies owning it, but not the 100+ licensees including Google.
Android is a key Linux distribution and Dalvik is an essential component of it. Also, someone else commented somewhere else on this thread that the Linux kernel itself is only a small percentage of all code of a Linux configuration.
OIN doesn't say it protects only the Linux kernel. It protects "the Linux System" according to its own definition. That definition is listed here. It's a long list of program files, and for an example, it includes Apache, which actually is available for Windows, too (unlike Dalvik), and which certainly isn't an operating system module.
That's what my criticism of the OIN is not exclusively but largely about: they have a very arbitrary definition of the scope of "protection". Why is Apache listed? Presumably because IBM, which is OIN's most important backer, has a strategic desire to protect Apache. If Apache is protected, I don't see why Dalvik shouldn't be. But OIN isn't about objective criteria. It's all about the strategic interests of six companies owning it, and plenty of others who become licensees are misled and misguided.
Just to make this clear, I don't mean to defend everything Google does or did, especially in connection with Android. Most Android-based phones appear to be closed source in practical terms, and the forking you mention plays a role in that.
But the OIN is not about free software or open source values. It claims to protect companies in the open source ecosystem, and Google became a licensee a couple of years ago and now sees that it doesn't get any benefit from its membership.
What IBM does with Websphere/Apache is also forking by the way.
SCO is a copyright case. While copyright litigation can also cause problems, there's a fundamental difference: you don't infringe copyright inadvertently. Theoretically you could, but practically you won't just by coincidence write a significant number of lines of code the same way someone else did. But patents are broad and you can infringe them totally unknowingly. That's why programmers who make independent creations never have to worry about copyright but unfortunately do have to worry about patents.
Concerning standardization, both Oracle and Google (as well as IBM and Red Hat) are member of "OpenForum Europe", a lobby group in the EU that pushes for "open standards". Here's a blog posting in which I criticized the hypocrisy of that group last month. If you look at the flawed Java Community Process, that's also a serious standardization problem.
All large corporations try to use the patent system or standardization processes and standards policy to their advantage...
It discussed many aspects of the dispute and in particular goes into detail on the seven patents Oracle seeks to enforce against Google, and inhowfar they may or may not read on Android.
I posted a correction in the comments there to point out that Oracle changed its stance on software patents years ago, not just after acquiring Sun's patent portfolio.
For years I've been criticizing all those fake solutions to the patent problem, such as "patent pledges" or the Open Invention Network (OIN). Both Google and Oracle are licensees of the OIN. The OIN patent agreement is meant to be a non-aggression pact between its members, with respect to "the Linux System".
Given that Android is a Linux distro (and a strategically very important one), it should be fully covered by the OIN as the self-proclaimed protective shield for the Linux ecosystem. Consequently, Oracle should be prohibited by the OIN cross-license agreement to sue its fellow OIN licensee Google. I'm not the only one to have raised that question. I saw Simon Phipps (OSI board member, former chief open source exec at Sun, now at ForgeRock) and Bradley Kuhn of the Software Freedom Law Center (and formerly FSF) raise the same kind of question on Twitter/identica. Now TheRegister contacted the OIN and wanted a comment on Oracle vs. Google, and the OIN declined to comment.
By the way, Eben Moglen promoted the OIN big time at LinuxCon, just a few days before Oracle announced its lawsuit.
What's certainly not a fake solution (although difficult to achieve) is the proposal to abolish software patents. The EndSoftPatents.org campaign runs the software patent wiki and has a pretty informative Wiki page on Oracle vs. Google.
"It's a sad comment on the morality of large modern software companies that Microsoft, while I don't think they've gotten any better since Sun sued them, probably has the high ground."
He seems to mean this primarily in terms of compliance with the official Java specification but one could also look at it in terms of software patent action against FOSS. I recently wrote about Microsoft's use of patents in connection with open source and got bashed for simply telling the truth: so far it's actually other companies who make the truly hostile moves. Far be it from me to defend software patents; I just mean to point out that there are different ways in which they get used, and in light of Oracle vs. Google, I believe more people will agree with me now.
One group of people James Gosling doesn't criticize are all those former Sun execs who strongly supported Oracle's acquisition of Sun because it was financially advantageous for them, only to leave the combined company as soon as possible after the closing of the deal. Many of those told the FOSS community that Sun had patents that could be very dangerous for open source, and Oracle was such a "reasonable" patent holder that it would be much better for them to acquire those patents than to take any risk that maybe Microsoft (which by the way never made a bid for Sun) could get them.
I opposed Oracle's acquisition of Sun. I also had concerns over the Java part but kept quiet about that and focused on MySQL. That's because I cooperated with Monty (the original author and founder of MySQL) and he wanted to be neutral about programming languages. For those who heard the slander that my work in that context aimed to change MySQL's license from the GPL to something else (which some even propagated here on Slashdot), I've meanwhile posted a detailed explanation, including links to several documents I used during my fight against the Oracle/Sun deal, in order to provide conclusive evidence that I argued against -- not for -- a license change. You can find that information in this blog posting (the link leads directly to a passage on MySQL and the GPL).
RMS is really a great visionary but Oracle actually proves him wrong because he recently warned against Mono, DotGNU and C# because of patent concerns. I disagreed and said that those platforms are the last pieces of software against Microsoft would consider using its patents because those basically help the.NET ecosystem. More importantly, I said that other programming languages are also patent-encumbered, and I mentioned Java. That's why I thought it was wrong to single out C#.
Oddly enough, right now -- just a few weeks later -- it looks like C#, which was submitted to a standards body, is actually much safer from a patent point of view than Java and much more of an open standard.
It will be hard to find out whether Oracle planned this kind of aggression when buying Sun, but it can certainly be stated that the free software/open source community hasn't benefited from the acquisition.
Did Oracle try to resolve this amicably with Google (by way of a license deal) or is Oracle pursuing purely destructive objectives?
Will Google solve this patent problem in a way that the entire Android ecosystem (including the makers of Android-based phones and the authors of Android apps) will be reassured, or will Google only take care of its own risk?
Is Java less of an open standard now than C#? I don't really buy the argument that Oracle may only be suing because of deviations from the standards definition. This kind of patent attack is evil no matter whether Google adhere to certain specififcations or not.
Isn't this now the ultimate proof that the Open Invention Network doesn't really protect the Linux ecosystem from patent attacks? This is case of one OIN licensee (Oracle) suing another (Google).
Where are those FOSS advocates who said that Oracle's acquisition of Sun would be good for the cause and for the community? Some of them even claimed that it was important to have Oracle acquire Sun's patents. I've documented that on my blog.
Is it perhaps time to forget about the community's favorite bogeyman and recognize that IBM, Oracle and others are a much more serious threat to FOSS at this stage?
How can the so-called OpenForum Europe lobby the European Union for open source/open standards when its two most powerful members, IBM and Oracle, are patent aggressors against open source, especially in interoperability contexts?
This is a patent dispute with very wide-ranging implications.
I agree with Bruce that Peer-to-Patent isn't the answer. In a way, it makes that system even stronger.
However, one cannot generally oppose the idea of patent quality. In other words, if there is a bad patent application and someone happens to spot it on Peer-to-Patent, there's nothing wrong with providing information that helps get the patent application rejected. I don't think we could argue that such patents should be granted, leaving it to someone else to seek their (costly) invalidation much later. A patent that can be invalidated but previously threatens many developers, distributors, publishers and users is a bad kind of pollution.
If the question is whether spending time on Peer-to-Patent is the best use of time to achieve improvement in connection with patents, the problem is that for the time being I don't see hands-down superior alternatives either for the time being. I still hope that the Defensive Patent License will create a framework under which one can obtain patents and use them against everyone except other DPL supporters. In that scenario, I believe it's better to spend time on applications for such patents and to assert them vigorously against those who refuse to support the DPL. There seems to be a delay with the finalization of the DPL, but I still believe it has huge potential. At least the DPL is company-independent.
At this stage the worst of all patent "defense" initiatives out there is the "Open" Invention Network (OIN). Before anyone spends time supporting that scheme, Peer-to-Patent is clearly the better choice.
One thing is that big corporations lobby for software patents all the time; another is that patent holders often treat inadvertent infringers as "pirates", mislabel them as "copycats" etc.
I'm all for intellectual property, but I think those patent holders who enforce their rights against third parties who've committed no wrongdoing often use inappropriate language in describing the "infringers". I understand the patent holders don't say, "we don't know whether the patents we own will even survive reexamination", but they're not just assertive: they also offend those who made independent creations that someone else just happened to patent before.
Let's be realistic and let's hope the local activists will keep an eye on developments: the abolition of software patents is far from certain in New Zealand, as I explained on my blog in detail.
To sum up the NZ problem quickly, the announcement made by the minister in charge talked about "a way forward for software patents", which isn't the same as doing away with them entirely, and indeed, the minister's announcement as well as the New Zealand Parliament's decision talk about allowing patents on devices with "embedded software".
My blog article discusses in detail that a distinction between "software" and "devices with embedded software" is extremely difficult under substantive patent law (the part of patent law that defines the scope of patentable subject matter). For an example, is an operating system of a computer "embedded software"? What about smartphones? Unfortunately, the New Zealand Parliament accepted the notion of patents on "telephones" with "embedded software", and while a small part of the operating software of such telephones is really telephone-specific, the largest part is just an operating system like any other.
Even a patent on a device with "embedded software" can be infringed by software: at least in the form of a contributory (indirect) infringement, which from the perspective of the software developer and publisher concerned comes down to the same exclusionary effect as a direct infringement.
To put it very simple, the problem is that the New Zealand Parliament and the minister in charge didn't say that they would do away with all software that could be infringed (be it on a direct or indirect/contributory basis) by software. They only talked about patents on software vs. patents on devices with embedded software, and even the latter can in principle be infringed by standalone software as well. It's harder to draw the line when coming from the angle of patentable subject matter than when looking at it from the infringement point of view. There's no guarantee in the law that software developers, publishers, distributors and users can't be sued over patents...
Again, for details on where this could go in terms of what patents they might allow or not, it's a complicated matter and I tried to explain it on my blog.
Show me in witting, on the IBM web site, where their software can not be run on mainframes built by other manufacturers. I don't want your blog, I want IBM official restrictions. Otherwise, your a troll. Nothing more, nothing less.
Read the European Commission's press release. You can be sure they read the z/OS licensing terms, and they have the expertise to interpret those.
OK, poor choice of words. I completely agree that PJ is impartial and focuses more on facts then anything else.
Do you think her presumed impartiality is corroborated by the fact that during all those years she never criticized anything IBM did, other than disagreeing with them on whether there should be software patents? That she claims to be against software patents but wrote that IBM is free to sue the pants off TurboHercules with patents? It's at the very least conspicuous that she's more loyal to IBM than Rush Limbaugh to the Republican Party.
Agreed, but what does that have to do with bundling software with hardware? Are you suggesting that people can imply some divide between two parts of the same product and then arbitrarily force a company to comply?
Not arbitrarily. There are criteria for that. In the EU, this isn't the first bundling issue: Microsoft was forced to offer Windows in a special version without the MediaPlayer, and last year an investigation concerning browsers was settled with the introduction of a browser choice dialog box. Ironically, IBM itself lobbied against Microsoft in those contexts.
Why is software and hardware sold together (both relying on each other) such a bad thing for a company to claim as a single product?
The bad thing is not to sell it together -- it's if customers who only want one are forced to buy the other as well, especially if one or both of them are market-dominating products.
There ARE alternatives to mainframes out there. There are alternatives to programming for one.
The mainframe business is unique, but especially if you have hugbe amounts of mission-critical legacy code, it's easier said than done that you could rewrite it all for some other system.
If companies don't like how IBM has been running things (virtually the same way decades, which is to say protectionist) then they should have been thinking of alternatives long ago instead of continually locking themselves into this situation. That old adage of, "Fool me once, shame on you. Fool me twice, shame on me." applies nicely here.
I believe I understand what you mean here, but there was a time when they didn't really have the choice you may be thinking of, and from an antitrust perspective, imperfect decision-making by customers must (within reason) be considered a market reality. That doesn't give IBM the right to squeeze those customers to no end.
You are the only one who has been pushing this hype while citing your own blog.
please go away
I discussed the mainframe antitrust issue here on slashdot when there was significant news: the NEON complaint against IBM, and today's announcement of the European Commission's investigation. I don't think an EU investigation is just "hype". That's very real.
Concerning my blog, I don't think the solution is for slashdot articles to be extremely long. At some point it makes sense to point to more detailed stories for those who care to read them. The source for the fact that the European Commission investigates is the EC's own website -- not my blog.
There's enough competition in the market for cars to take care of that problem. But there isn't enough competition for mainframes. There isn't any significant competitive pressure on IBM in that field.
It's not illegal to have a monopoly. It's not illegal to bundle hardware and software. The combination of a dominant market position and abusive conduct is, however, an antitrust violation.
The fact that the European Commission has opened an investigation shows that there are serious issues involved. The launch of the investigation isn't a final ruling, but that type of decision isn't taken unless there's a lot of hard facts on the table.
So, as a minor side effect, today's announcement calls into question the accurancy and relevance of the things PJ claimed. She made TurboHercules out to be unreasonable. But TurboHercules just asked IBM for something that's in the interest of competition and innovation, and if the Commission viewed it differently (or if the Commission viewed this as simply a Microsoft initiative), the full-blown investigation wouldn't have been launched today.
As an alternative question, who would really want to "steal" z/OS anyway.
TurboHercules just asked for a fair, reasonable and non-discriminatory licensing offer. In other words, customers should pay for it. They just shouldn't be forced to run it exclusively on IBM hardware.
Does it even run on anything other an an IBM mainframe?
Yes. Emulation is the answer.
And if you have the money for the hardware, chances are you want the software... or maybe Linux.
You can run GNU/Linux on a mainframe, it's called z/Linux. However, if you have mission-critical programs running on z/OS, you're locked in right now and porting is easier said than done. There are estimates that a couple hundred billion lines of legacy code is still in use on mainframes (by banks, insurance companies, social security, etc.).
If IBM can't disallow the use of z/OS under an emulator, does it also mean that Apple can't disallow OS X on Hackintoshes?
You already got a pretty accurate answer from 'betterunixthanunix' about market share. I also explained that on a different branch of this discussion tree, here. Concerning Apple, they may be affected by a different EU initiative further down the road.
Wow...a company tying it's software to the hardware they manufactured! Holy crap...who could have imagined!?!
IBM isn't the only company doing that. But they have a de facto monopoly on mainframes and that's why there's an antitrust issue. It takes a market position that is at least dominant (IBM is even superdominant on mainframes) AND anticompetitive behavior. One of the two isn't enough to make a legal case. I discussed the question of market dominance versus significance in this recent blog posting.
Diversion is not a defense -- especially if it doesn't make sense.
Microsoft had to pay the highest antitrust fine in EU history and was pursued over three antitrust issues in recent years (MediaPlayer, Samba, browsers; MediaPlayer and Samba were one case but with two distinct parts). It doesn't look like the Commission does Microsoft's bidding;-)
The Commission's press release also mentions an issue about maintenance. No company whatsoever filed a complaint about that. But the Commission, probably because it looked into the mainframe market as a whole as a result of those different complaints, apparently determined that there's a problem that needs further scrutiny.
Hercules is an 11-year old open source project. Did Microsoft start that one back in 1999 in order to hurt IBM a decade later?
There's also the NEON complaint. NEON's principal founder is a billionaire and former co-founder of BMC, a big enterprise/mainframe software company.
The bottom line is that IBM will have to address the concerns identified -- on substance, not on the basis of conspiracy theories.
Slashdot Mirror
There's a fundamental error in the InfoWorld analysis referenced above:
Oracle simply asks for absolutely standard remedies in this situation. There's nothing evil about it, and it cannot be reasonably interpreted as a strategy to destroy open source as a whole or anything like that.
I'm saying this even though I opposed Oracle's acquisition of Sun. I just want to point out that if a case like this goes to court, the plaintiff will always ask for those kinds of remedies. There's nothing unusual about it. In fact, asking for less would be unusual and would probably confuse the judges as to what Oracle actually wants.
Intellectual property rights are exclusive rights. That's the way the law has designed them -- it's not a matter of Oracle being evil. Those IPRs entitle a right holder to enforce exclusivity. That necessarily means to ask for an injunction, and under such circumstances as the ones of this case (with copyright in play), also the destruction of infringing material.
The way to prevent that scenario from materializing is a license agreement between Oracle and Google. So it's up to the two parties to sit down and negotiate, and I believe we as a community should now expect both of them to be constructive. The court can't impose a license agreement on the two of them. If the court has to rule, it will -- if Oracle is right -- have to enforce exclusivity. That's sort of binary, whereas a license agreement would offer much more flexibility.
It's regrettable that they couldn't work this out before the matter was taken to court. But it's not too late until there is a final court ruling.
I just posted further up an explanation as to why the GPLv2's implicit patent grant doesn't really help forks.
There's some confusion out there about how Google would be in a better position from a patent point of view if it had used existing Java code under the GPL (OpenJDK, phoneME). The Android ecosystem as a whole would have had other benefits (such as making it much harder for the makers of Android-based phones to keep important parts of their source code closed) but it wouldn't really help as far as Oracle's patent infringement allegations are concerned.
The GPLv2 (under which OpenJDK and large parts of phoneME are available) does not contain an explicit patent grant. Only an implicit one. As a result, any fork (derived/modified version) is probably not covered.
The InfoWorld article that this Slashdot story refers to talks a lot about forking as a possible strategy -- especially toward the end, where MySQL is also mentioned. I was very much involved with the debate over whether Oracle should get to acquire MySQL (together with Monty, MySQL's original author/founder, I opposed the deal). In that context, it was also a subject of debate whether MySQL forks would be safe from Oracle patent threats in the future. Eben Moglen, who was basically part of Oracle's legal team and had botched the patent aspect of GPLv2 (thus tried to fix the problem with GPLv3), argued that GPLv2 would take care of those forks. However, the European Commission, which (unlike Moglen) is impartial and has vast legal resources, concluded that the implicit patent grant does not -- at least not reliably, but probably not at all -- protect forks.
If you're interested in more detail on the question of whether Google would be or would have been better off with GPLv2, here's a link to the related part of a blog posting of mine. It discusses that question and right thereafter (or you can go there directly) explains that my work related to Oracle's acquisition of MySQL was not an effort to change MySQL's license away from the GPL to something else. I have meanwhile published documents from the process that serve as conclusive evidence that I argued vehemently against -- not for -- a license change. Still, the GPLv2's limitation concerning patent claims against forks is a fact.
I commented on this elsewhere in this discussion: http://yro.slashdot.org/comments.pl?sid=1756418&cid=33275698
Indeed, Apache is covered, and so are lots of other things that go way beyond the Linux kernel. The problem is that the OIN has that completely arbitrary approach to determining what's on the list and what's not. I actually suggested four alternative ways to address that problem, but it seems they don't want to change anything. They want the OIN to continue to be totally intransparent (which in terms of litigation would be understandable, but the definition of the scope could be a transparent process). And they want it to simply benefit the six companies owning it, but not the 100+ licensees including Google.
Android is a key Linux distribution and Dalvik is an essential component of it. Also, someone else commented somewhere else on this thread that the Linux kernel itself is only a small percentage of all code of a Linux configuration.
OIN doesn't say it protects only the Linux kernel. It protects "the Linux System" according to its own definition. That definition is listed here. It's a long list of program files, and for an example, it includes Apache, which actually is available for Windows, too (unlike Dalvik), and which certainly isn't an operating system module.
That's what my criticism of the OIN is not exclusively but largely about: they have a very arbitrary definition of the scope of "protection". Why is Apache listed? Presumably because IBM, which is OIN's most important backer, has a strategic desire to protect Apache. If Apache is protected, I don't see why Dalvik shouldn't be. But OIN isn't about objective criteria. It's all about the strategic interests of six companies owning it, and plenty of others who become licensees are misled and misguided.
Just to make this clear, I don't mean to defend everything Google does or did, especially in connection with Android. Most Android-based phones appear to be closed source in practical terms, and the forking you mention plays a role in that.
But the OIN is not about free software or open source values. It claims to protect companies in the open source ecosystem, and Google became a licensee a couple of years ago and now sees that it doesn't get any benefit from its membership.
What IBM does with Websphere/Apache is also forking by the way.
SCO is a copyright case. While copyright litigation can also cause problems, there's a fundamental difference: you don't infringe copyright inadvertently. Theoretically you could, but practically you won't just by coincidence write a significant number of lines of code the same way someone else did. But patents are broad and you can infringe them totally unknowingly. That's why programmers who make independent creations never have to worry about copyright but unfortunately do have to worry about patents.
Concerning standardization, both Oracle and Google (as well as IBM and Red Hat) are member of "OpenForum Europe", a lobby group in the EU that pushes for "open standards". Here's a blog posting in which I criticized the hypocrisy of that group last month. If you look at the flawed Java Community Process, that's also a serious standardization problem.
All large corporations try to use the patent system or standardization processes and standards policy to their advantage...
I wanted to recommend this detailed blog posting (about 8,500 words, plenty of scrolling) on Oracle vs. Google.
It discussed many aspects of the dispute and in particular goes into detail on the seven patents Oracle seeks to enforce against Google, and inhowfar they may or may not read on Android.
I posted a correction in the comments there to point out that Oracle changed its stance on software patents years ago, not just after acquiring Sun's patent portfolio.
For years I've been criticizing all those fake solutions to the patent problem, such as "patent pledges" or the Open Invention Network (OIN). Both Google and Oracle are licensees of the OIN. The OIN patent agreement is meant to be a non-aggression pact between its members, with respect to "the Linux System".
Given that Android is a Linux distro (and a strategically very important one), it should be fully covered by the OIN as the self-proclaimed protective shield for the Linux ecosystem. Consequently, Oracle should be prohibited by the OIN cross-license agreement to sue its fellow OIN licensee Google. I'm not the only one to have raised that question. I saw Simon Phipps (OSI board member, former chief open source exec at Sun, now at ForgeRock) and Bradley Kuhn of the Software Freedom Law Center (and formerly FSF) raise the same kind of question on Twitter/identica. Now TheRegister contacted the OIN and wanted a comment on Oracle vs. Google, and the OIN declined to comment.
By the way, Eben Moglen promoted the OIN big time at LinuxCon, just a few days before Oracle announced its lawsuit.
What's certainly not a fake solution (although difficult to achieve) is the proposal to abolish software patents. The EndSoftPatents.org campaign runs the software patent wiki and has a pretty informative Wiki page on Oracle vs. Google.
He seems to mean this primarily in terms of compliance with the official Java specification but one could also look at it in terms of software patent action against FOSS. I recently wrote about Microsoft's use of patents in connection with open source and got bashed for simply telling the truth: so far it's actually other companies who make the truly hostile moves. Far be it from me to defend software patents; I just mean to point out that there are different ways in which they get used, and in light of Oracle vs. Google, I believe more people will agree with me now.
One group of people James Gosling doesn't criticize are all those former Sun execs who strongly supported Oracle's acquisition of Sun because it was financially advantageous for them, only to leave the combined company as soon as possible after the closing of the deal. Many of those told the FOSS community that Sun had patents that could be very dangerous for open source, and Oracle was such a "reasonable" patent holder that it would be much better for them to acquire those patents than to take any risk that maybe Microsoft (which by the way never made a bid for Sun) could get them.
I opposed Oracle's acquisition of Sun. I also had concerns over the Java part but kept quiet about that and focused on MySQL. That's because I cooperated with Monty (the original author and founder of MySQL) and he wanted to be neutral about programming languages. For those who heard the slander that my work in that context aimed to change MySQL's license from the GPL to something else (which some even propagated here on Slashdot), I've meanwhile posted a detailed explanation, including links to several documents I used during my fight against the Oracle/Sun deal, in order to provide conclusive evidence that I argued against -- not for -- a license change. You can find that information in this blog posting (the link leads directly to a passage on MySQL and the GPL).
RMS is really a great visionary but Oracle actually proves him wrong because he recently warned against Mono, DotGNU and C# because of patent concerns. I disagreed and said that those platforms are the last pieces of software against Microsoft would consider using its patents because those basically help the .NET ecosystem. More importantly, I said that other programming languages are also patent-encumbered, and I mentioned Java. That's why I thought it was wrong to single out C#.
Oddly enough, right now -- just a few weeks later -- it looks like C#, which was submitted to a standards body, is actually much safer from a patent point of view than Java and much more of an open standard.
It will be hard to find out whether Oracle planned this kind of aggression when buying Sun, but it can certainly be stated that the free software/open source community hasn't benefited from the acquisition.
There's a number of important questions that Oracle's patent attack raises:
This is a patent dispute with very wide-ranging implications.
I agree with Bruce that Peer-to-Patent isn't the answer. In a way, it makes that system even stronger.
However, one cannot generally oppose the idea of patent quality. In other words, if there is a bad patent application and someone happens to spot it on Peer-to-Patent, there's nothing wrong with providing information that helps get the patent application rejected. I don't think we could argue that such patents should be granted, leaving it to someone else to seek their (costly) invalidation much later. A patent that can be invalidated but previously threatens many developers, distributors, publishers and users is a bad kind of pollution.
If the question is whether spending time on Peer-to-Patent is the best use of time to achieve improvement in connection with patents, the problem is that for the time being I don't see hands-down superior alternatives either for the time being. I still hope that the Defensive Patent License will create a framework under which one can obtain patents and use them against everyone except other DPL supporters. In that scenario, I believe it's better to spend time on applications for such patents and to assert them vigorously against those who refuse to support the DPL. There seems to be a delay with the finalization of the DPL, but I still believe it has huge potential. At least the DPL is company-independent.
At this stage the worst of all patent "defense" initiatives out there is the "Open" Invention Network (OIN). Before anyone spends time supporting that scheme, Peer-to-Patent is clearly the better choice.
One thing is that big corporations lobby for software patents all the time; another is that patent holders often treat inadvertent infringers as "pirates", mislabel them as "copycats" etc.
I'm all for intellectual property, but I think those patent holders who enforce their rights against third parties who've committed no wrongdoing often use inappropriate language in describing the "infringers". I understand the patent holders don't say, "we don't know whether the patents we own will even survive reexamination", but they're not just assertive: they also offend those who made independent creations that someone else just happened to patent before.
IBM's statements concerning some mainframe-related technologies, including the Hercules open source mainframe emulator, are particularly obnoxious.
Let's be realistic and let's hope the local activists will keep an eye on developments: the abolition of software patents is far from certain in New Zealand, as I explained on my blog in detail.
To sum up the NZ problem quickly, the announcement made by the minister in charge talked about "a way forward for software patents", which isn't the same as doing away with them entirely, and indeed, the minister's announcement as well as the New Zealand Parliament's decision talk about allowing patents on devices with "embedded software".
My blog article discusses in detail that a distinction between "software" and "devices with embedded software" is extremely difficult under substantive patent law (the part of patent law that defines the scope of patentable subject matter). For an example, is an operating system of a computer "embedded software"? What about smartphones? Unfortunately, the New Zealand Parliament accepted the notion of patents on "telephones" with "embedded software", and while a small part of the operating software of such telephones is really telephone-specific, the largest part is just an operating system like any other.
Even a patent on a device with "embedded software" can be infringed by software: at least in the form of a contributory (indirect) infringement, which from the perspective of the software developer and publisher concerned comes down to the same exclusionary effect as a direct infringement.
To put it very simple, the problem is that the New Zealand Parliament and the minister in charge didn't say that they would do away with all software that could be infringed (be it on a direct or indirect/contributory basis) by software. They only talked about patents on software vs. patents on devices with embedded software, and even the latter can in principle be infringed by standalone software as well. It's harder to draw the line when coming from the angle of patentable subject matter than when looking at it from the infringement point of view. There's no guarantee in the law that software developers, publishers, distributors and users can't be sued over patents...
Again, for details on where this could go in terms of what patents they might allow or not, it's a complicated matter and I tried to explain it on my blog.
Show me in witting, on the IBM web site, where their software can not be run on mainframes built by other manufacturers. I don't want your blog, I want IBM official restrictions. Otherwise, your a troll. Nothing more, nothing less.
Read the European Commission's press release. You can be sure they read the z/OS licensing terms, and they have the expertise to interpret those.
OK, poor choice of words. I completely agree that PJ is impartial and focuses more on facts then anything else.
Do you think her presumed impartiality is corroborated by the fact that during all those years she never criticized anything IBM did, other than disagreeing with them on whether there should be software patents? That she claims to be against software patents but wrote that IBM is free to sue the pants off TurboHercules with patents? It's at the very least conspicuous that she's more loyal to IBM than Rush Limbaugh to the Republican Party.
Agreed, but what does that have to do with bundling software with hardware? Are you suggesting that people can imply some divide between two parts of the same product and then arbitrarily force a company to comply?
Not arbitrarily. There are criteria for that. In the EU, this isn't the first bundling issue: Microsoft was forced to offer Windows in a special version without the MediaPlayer, and last year an investigation concerning browsers was settled with the introduction of a browser choice dialog box. Ironically, IBM itself lobbied against Microsoft in those contexts.
Why is software and hardware sold together (both relying on each other) such a bad thing for a company to claim as a single product?
The bad thing is not to sell it together -- it's if customers who only want one are forced to buy the other as well, especially if one or both of them are market-dominating products.
There ARE alternatives to mainframes out there. There are alternatives to programming for one.
The mainframe business is unique, but especially if you have hugbe amounts of mission-critical legacy code, it's easier said than done that you could rewrite it all for some other system.
If companies don't like how IBM has been running things (virtually the same way decades, which is to say protectionist) then they should have been thinking of alternatives long ago instead of continually locking themselves into this situation. That old adage of, "Fool me once, shame on you. Fool me twice, shame on me." applies nicely here.
I believe I understand what you mean here, but there was a time when they didn't really have the choice you may be thinking of, and from an antitrust perspective, imperfect decision-making by customers must (within reason) be considered a market reality. That doesn't give IBM the right to squeeze those customers to no end.
You are the only one who has been pushing this hype while citing your own blog.
please go away
I discussed the mainframe antitrust issue here on slashdot when there was significant news: the NEON complaint against IBM, and today's announcement of the European Commission's investigation. I don't think an EU investigation is just "hype". That's very real.
Concerning my blog, I don't think the solution is for slashdot articles to be extremely long. At some point it makes sense to point to more detailed stories for those who care to read them. The source for the fact that the European Commission investigates is the EC's own website -- not my blog.
There's enough competition in the market for cars to take care of that problem. But there isn't enough competition for mainframes. There isn't any significant competitive pressure on IBM in that field.
It's not illegal to have a monopoly. It's not illegal to bundle hardware and software. The combination of a dominant market position and abusive conduct is, however, an antitrust violation.
The fact that the European Commission has opened an investigation shows that there are serious issues involved. The launch of the investigation isn't a final ruling, but that type of decision isn't taken unless there's a lot of hard facts on the table.
So, as a minor side effect, today's announcement calls into question the accurancy and relevance of the things PJ claimed. She made TurboHercules out to be unreasonable. But TurboHercules just asked IBM for something that's in the interest of competition and innovation, and if the Commission viewed it differently (or if the Commission viewed this as simply a Microsoft initiative), the full-blown investigation wouldn't have been launched today.
As an alternative question, who would really want to "steal" z/OS anyway.
TurboHercules just asked for a fair, reasonable and non-discriminatory licensing offer. In other words, customers should pay for it. They just shouldn't be forced to run it exclusively on IBM hardware.
Does it even run on anything other an an IBM mainframe?
Yes. Emulation is the answer.
And if you have the money for the hardware, chances are you want the software... or maybe Linux.
You can run GNU/Linux on a mainframe, it's called z/Linux. However, if you have mission-critical programs running on z/OS, you're locked in right now and porting is easier said than done. There are estimates that a couple hundred billion lines of legacy code is still in use on mainframes (by banks, insurance companies, social security, etc.).
If IBM can't disallow the use of z/OS under an emulator, does it also mean that Apple can't disallow OS X on Hackintoshes?
You already got a pretty accurate answer from 'betterunixthanunix' about market share. I also explained that on a different branch of this discussion tree, here. Concerning Apple, they may be affected by a different EU initiative further down the road.
Wow...a company tying it's software to the hardware they manufactured! Holy crap...who could have imagined!?!
IBM isn't the only company doing that. But they have a de facto monopoly on mainframes and that's why there's an antitrust issue. It takes a market position that is at least dominant (IBM is even superdominant on mainframes) AND anticompetitive behavior. One of the two isn't enough to make a legal case. I discussed the question of market dominance versus significance in this recent blog posting.
Diversion is not a defense -- especially if it doesn't make sense.
Microsoft had to pay the highest antitrust fine in EU history and was pursued over three antitrust issues in recent years (MediaPlayer, Samba, browsers; MediaPlayer and Samba were one case but with two distinct parts). It doesn't look like the Commission does Microsoft's bidding ;-)
The Commission's press release also mentions an issue about maintenance. No company whatsoever filed a complaint about that. But the Commission, probably because it looked into the mainframe market as a whole as a result of those different complaints, apparently determined that there's a problem that needs further scrutiny.
Hercules is an 11-year old open source project. Did Microsoft start that one back in 1999 in order to hurt IBM a decade later?
There's also the NEON complaint. NEON's principal founder is a billionaire and former co-founder of BMC, a big enterprise/mainframe software company.
The bottom line is that IBM will have to address the concerns identified -- on substance, not on the basis of conspiracy theories.