Sun's patent covenant applies to any version of the ODF specification in which Sun participates. Subsequent specifications built on those versions would also benefit from the covenant to the extent that they are the same as the earlier versions. Only changes introduced by the new version would be excluded. That's the sense in which the covenant was framed by Sun (I was involved), and in which it has been understood by the reasonable people who have read it to date. Analysis by independent counsel raised no such issue - do you have independent counter-advice?
Your argument is based on a twisted interpretation of the Sun covenant, and your inflammatory and alarmist language about "holding to hostage" is incorrect. Sun has no intention of leaving the OASIS OpenDocument TC anyway, so in addition it's intensely hypothetical. I note you argued this same point before, at length. Please stop. You are wrong, and I am actually in a position to assert that.
A kind of GPL for standards indemnification against patents that leaves no questions open and is applicable on a worldwide scale?
I agree completely and I approached OSI at OSCON with a view to OSI creating a definition of just such a patent covenant (having Sun create it unilaterally is unlikely to be welcomed by IBM and Microsoft!). My first attempt at an outline was in my blog posting on the subject and I hope the OSI Board will commission a group next month.
Thanks for the pointer, appreciated (for others looking, it's in this comment). That's actually how the ODF plug-in achieves the addition to the format subsystem. It works as well as any of the other import/export capabilities that use this mechanism do (read: all of them). I struggle to understand why this level of interoperability hasn't been added directly to MS Office rather than Sun having to create a plug-in.
Extending the format system using this approach achieves the goal our team was trying to achieve, which was to ensure that users of assistive devices would be able to access ODF as part of their work flow within Word. They would be unable to switch immediately to OpenOffice.org because of the tight-coupling of the assistive devices to MS-Office by their manufacturers. Brian's goal (importing ODF directly into OOXML in order to encourage use of OOXML) is different and he's probably right that it's harder to achieve - sadly their chosen mechanism is more of a barrier to disabled customers.
Which says that MS themselves said that you cannot change the default 'save as'.
Given that Sun's ODF plug-in for MS Word does exactly that (simply adds ODF to the list of supported formats everywhere they occur in Word, including allowing it to be set as the default), it makes me wonder why Microsoft would say that.
b) Non-discriminatory. This means that you have to license to all comers. You can't decide you like what this company is doing but not this other company. Anyone who pays the moneys get the licenses.
Focussing just on the money can easily make one miss the deeper problem with RAND terms. You'll find that some people now talk about RANDZ license terms, where the "Z" indicates "license fee is $0". But that doesn't necessarily make them any better. The license also has to allow sub-licensing without referral to the licensor. If it doesn't, open source implementations are effectively impossible since every developer and every user will need to independently obtain a license from the patent holder - an impossibly chilling effect.
Microsoft covenant inferior to Sun covenant
on
Patent Threats In OOXML
·
· Score: 4, Informative
I don't normally respond to trolls, but I want to make sure this is clear. Despite the claims of Microsoft's representatives, their patent covenant is not the same as Sun's. There are several important differences, as I pointed out at the time:
Microsoft make their promise contingent on the patents being "essential", at their sole judgement, to the implementation involved. There may be several ways to implement each feature; if you happen to pick the one covered by the patent, you are using one that's not "essential" since you could be implementing one of the alternative ways. You can't know this without extensive research and legal advice.
They also make it contingent on "conformance", again at their sole discretion. Partial implementations may be at risk, and since open source development is done in public, so may in-progress full implementations.
Thirdly, despite placing these limitations on their outward grant, they expect all recipients of the grants to refrain from all litigation, not just that bounded by either conformance or essential claims.
Items 1 and 2 are especially important. By reserving unaccountable judgement over what is and is not covered, they prevent implementors having certainty they will not face patent issues. This is exactly the way to chill the enthusiasm of open source developers, for whom certainty over their freedoms is the cornerstone of community. It's exactly the reason I made sure Sun's covenant was not crippled in the same way.
I have now had several reports of Microsoft's representatives claiming their covenant is the same as Sun's; it is not, please make sure anyone who says so is challenged.
There's one more issue of note, which the NZ paper makes clear. Microsoft explicitly uses proprietary formats within their MS-OOXML specification (DrawingML for example). If they want to provide comfort to open source developers, they need to go further and cover all referenced formats with their "promise" as well.
The original T1 design was released under GPLv2 a while back, and the new T2 design will be released under GPLv2 for now as well. Using GPL v2 creates an implicit patent license so while using GPLv3 would make the situation cleared GPLv2 is probably sufficient.
And yes if you look at the map on the opensparc.net page (when they get some quota back after being slashdotted) you'll see they are getting a vast amount of interest from China, where I gather a company is already producing an OpenSPARC T1-derived chip for embedded use.
The plug-in includes some code that inserts the file format as an option in the "File Open/Save" dialog and in the rest of the file handling for MS Office so that the ODF formats appear as natural parts of Office (rather than in the un-natural way that the plug-in on SourceForge that MS sponsor tries to work). I gather that the developers are not sure if they can open that small piece of source code so the package as a whole is not under an open source license yet. The rest of the plug-in (the vast majority) is OpenOffice.org code and is open source.
These two I'd expect, but I'd also have expected this work to be well under-way by now, and hopefully nearing completion. I read about Sun being interested in GPL3 almost 18 months ago, and I would have thought a review and documentation process would have begun then, if not earlier.
Right - I've got a good handle on the license overall. However, there were some very significant changes in the last draft, final call and final version which may have significant consequences in our business (not least with how we are able to defend ourselves against patent aggressors) and we'll need to do some deep thinking about them.
Even if GPL3 winds up being wrong for Sun (I hope not for the sake of Sun and the sake of the community) at some point it's just good corporate governance to have a handle on who owns your code.
Oh, we know where every line of code that's been Freed came from, have no fear! The engineering teams for both OpenSolaris and OpenJDK did a phenomenal job of source code archaeology. No, my point was that changing the file headers is actually a non-trivial task that triggers a wave of regression testing. Not to be undertaken lightly.
I appreciate that you may be restricted on what you can say about this (though open-ness is a good goal) but if Sun is seen to be waiting for GPL3 final text before beginning such a process it would be seen as much less serious.
Yep, we were waiting for the final text. We engaged fully, contributed throughout but the final text was still news to us at the end of the week. And even when due diligence is done, we've still the need to get approval from the communities around the code in each case. I remain positive though, and I'll be trying hard to stay transparent (though some things just have to be kept internal of course).
You may be interested in the commentary I have written explaining Sun's position on GPLv3. Despite having been involved in the discussion process from the beginning, we still only saw the final text at the end of this week. Whereas some people (who ought to know better) think we can just change the licenses over-night, in fact we have to take into account:
The meaning of the final text and what its implications are. The team has started the discussion but with July 4 week next week many of them will be taking vacation so I don't expect a firm statement for several weeks.
The fact (which I explain at length in my blog) that Sun is only the steward of the copyrights and the decision to change licenses takes more agreement than just Sun's.
The fact that we've 6+ million lines of code in each of OpenSolaris (which is under CDDL) and OpenJDK (which is under GPLv2) and it will be heavy lifting to make a change.
So please do let Jonathan know how you feel, but be aware he's not the only guy who has to decide.
Actually, Sun has promised to release the Java platform under GPLv2, not GPLv3. I'll review GPLv3 when it comes out, though.
I want the pressure kept on them to live up to their promise to do so.
Sun will release the Java SE class libraries under GPLv2 (less the encumbered portions previously explained) before the end of June as promised.
Ubuntu's inclusion of a non-Free version relieves that pressure.
In fact, the Java software is in Multiverse so has the same status as the JDK included in Debian's non-Free repository. Please don't ding Canonical this way. They have done more than you can possibly imagine to help Sun's staff with the work. They have worked faithfully with Sun's executives, persuading them of the need for Free software. They keep the pressure up by co-operating and setting the pace. They deserve your praise and not this criticism.
It's a bug, sorry - a placeholder installer was used during testing because of a licensing issue that needed sorting, and it didn't get replaced in time with the real.deb - we're working on it.
Thank you for illustrating my point again that we can't go by press releases or zealots like you in making platform decisions, neither with Solaris nor with Java.
OK, I'm now feeding the troll, but I'll not let that smear stand. Sun has been scrupulously clear over exactly what is and is not under open source licenses. There are indeed parts of Solaris that are still encumbered, but that doesn't stop Nexenta and three other distributions building complete distributions, any more than the use of binary drivers by varying GNU/Linux distros invalidates the use of the term "Free" to describe them unless you are a complete radical - oh, sorry, it's me that's the zealot, I forgot.
The full source to the Java class libraries will be released in the spring under GNU/Classpath apart from the bits that hostile parties refuse to relicense to Sun. Fortunately the Classpath folks are helping work round that. Now, how about some working-with rather than all this negativity?
Well, firstly, Solaris is certainly open source, Really? All of it? Or just some of it? Who can even tell.
If you are actually interested and not trolling, I'd suggest you join the other 18,000 or so people over on OpenSolaris.org where the source code used to build Solaris is available under a Free license. You'll find a number of other OS distributions based on the same code - the most interesting is probably NextentaOS, which is essentially Debian with the kernel switched.
No, I got it, and I disagree with you. My amazement stems from the fact that some very smart people who I respect (the Linux kernel developers) are criticising the GPLv3 for things the discussion committees and the FSF are already addressing. I actually think that's clear from the original blog posting.
"Sun has licensed Unix System V from Santa Cruz Operation for several years now..." No. SUN has 100% autonomous rights to Solaris, like IBM has 100% autonomous rights to AIX, and hasn't owed royalty payments to anyone for many years. SUN paid SCO millions of dollars that SUN was not obligated to pay, and to which SCO was not entitled to receive. At the time, several SUN executives were on an anti-Linux, anti-GPL, anti-IBM rampage. It was widely seen (and I agree 100%) as SUN giving SCO money to harm Linux.
There's even more to it than that actually. First, the rights Sun has to Unix are far more extensive than those IBM has to AIX. In 1994, Sun bought an outright perpetual copyright license giving it non-exclusive ownership of the Unix source code. That's the basis on which Sun has been able to open source Unix in the OpenSolaris community. Sun has been fastidious in constantly acquiring all rights to Unix so that there could never be another party with a claim over Sun's core business.
However, that doesn't mean Sun (or any other Unix licensee) has had constant access to updates and new features. Whereas in the world of open source we all share enhancements with each other, in the world of Unix all the licensees retained their enhancements. The Unix copyright owner (originally AT&T, then Novell, then SCO) had the practice of buying enhancements (including drivers) from the licensees and then selling then to other licensees. This "clearing-house" role was what allowed Unix features to flow between the licensees without an open source community.
What Sun and all the rest of the licensees did was trade code with the Unix Licensor, usually for a net payment to the Licensor. That last payment was a confidential transaction like all the previous ones made by Sun, IBM, HP and others, but was a pretty routine transaction to secure updates and maintain a full license to the Solaris source. While I recognise that plenty of folks view that last transaction as "funding SCO", it was at its heart unexceptional apart from being the final rights acquisition before open sourcing the whole code-base. And having seen how SCO behaved since then, it was the last.
Re:Thought the summary was vacuous...
on
Sun Exec Backs GPLv3
·
· Score: 3, Informative
It's a hot topic and there are plenty of voices for and against. It was like this discussing the use of the GPL for the Java platform though, extensive and passionate debate right up to the last minute. In the end GPL v2 with the Classpath exception was clearly preferable, but every possible option was explored.
One key difference with Solaris though is that the base source code is already open source Free software and is in the care of the OpenSolaris community, so while Sun obviously gets a big say in what happens it's not just down to us, the community will also need to discuss it.
Re:Thought the summary was vacuous...
on
Sun Exec Backs GPLv3
·
· Score: 2, Informative
In fact my amazement stems from the fact that intelligent people who I respect are criticising things (like the DRM language) that have already been substantially altered, but arguing as if there has been no change to them. Sorry you find the whole thing "vacuous", you are clearly a Higher Intellect.
To see what I actually said, rather than relying on the ZDNet extracts, you'll find the original blog posting here. It's linked from the article but clearly from some of the comments above some people haven't spotted the link.
So, it's up to the courts to sort out the question of prior art. The common complaint here on slashdot (and in my own heart) is that the USPTO should check prior art... but now that I think about it, that seems inefficient. It's inefficient because most patents will never come to dispute, and hence are irrelevant. Why spend a gabillion dollars bringing in the necessary expertise, until you know that it actually matters (i.e. is actually challenged)?
Seems to me that the main problem here is that there's no meaningful penalty for ignoring prior art. One idea that seems easy but I've never seen pursued is for the law to be changed to treat a failure to cite prior art as perjury. Then, should a successful prior art case be prosecuted against a patent, the applicant would be subject to fines or even imprisonment. This simple change would rebalance the system and result in far fewer lame patents with obvious prior art.
The page ostensibly lists all licenses used in Debian, and has no mention of the CDDL. Personally, I was able to infer from this that the CDDL is not an approved license
I suggest it would be wiser to infer that no package has so far been included in Debian with that license. There are plenty of licenses missing from that list.
Slashdot Mirror
Sun's patent covenant applies to any version of the ODF specification in which Sun participates. Subsequent specifications built on those versions would also benefit from the covenant to the extent that they are the same as the earlier versions. Only changes introduced by the new version would be excluded. That's the sense in which the covenant was framed by Sun (I was involved), and in which it has been understood by the reasonable people who have read it to date. Analysis by independent counsel raised no such issue - do you have independent counter-advice?
Your argument is based on a twisted interpretation of the Sun covenant, and your inflammatory and alarmist language about "holding to hostage" is incorrect. Sun has no intention of leaving the OASIS OpenDocument TC anyway, so in addition it's intensely hypothetical. I note you argued this same point before, at length. Please stop. You are wrong, and I am actually in a position to assert that.
I agree completely and I approached OSI at OSCON with a view to OSI creating a definition of just such a patent covenant (having Sun create it unilaterally is unlikely to be welcomed by IBM and Microsoft!). My first attempt at an outline was in my blog posting on the subject and I hope the OSI Board will commission a group next month.
Thanks for the pointer, appreciated (for others looking, it's in this comment). That's actually how the ODF plug-in achieves the addition to the format subsystem. It works as well as any of the other import/export capabilities that use this mechanism do (read: all of them). I struggle to understand why this level of interoperability hasn't been added directly to MS Office rather than Sun having to create a plug-in.
Extending the format system using this approach achieves the goal our team was trying to achieve, which was to ensure that users of assistive devices would be able to access ODF as part of their work flow within Word. They would be unable to switch immediately to OpenOffice.org because of the tight-coupling of the assistive devices to MS-Office by their manufacturers. Brian's goal (importing ODF directly into OOXML in order to encourage use of OOXML) is different and he's probably right that it's harder to achieve - sadly their chosen mechanism is more of a barrier to disabled customers.
Given that Sun's ODF plug-in for MS Word does exactly that (simply adds ODF to the list of supported formats everywhere they occur in Word, including allowing it to be set as the default), it makes me wonder why Microsoft would say that.
Focussing just on the money can easily make one miss the deeper problem with RAND terms. You'll find that some people now talk about RANDZ license terms, where the "Z" indicates "license fee is $0". But that doesn't necessarily make them any better. The license also has to allow sub-licensing without referral to the licensor. If it doesn't, open source implementations are effectively impossible since every developer and every user will need to independently obtain a license from the patent holder - an impossibly chilling effect.
I don't normally respond to trolls, but I want to make sure this is clear. Despite the claims of Microsoft's representatives, their patent covenant is not the same as Sun's. There are several important differences, as I pointed out at the time:
Items 1 and 2 are especially important. By reserving unaccountable judgement over what is and is not covered, they prevent implementors having certainty they will not face patent issues. This is exactly the way to chill the enthusiasm of open source developers, for whom certainty over their freedoms is the cornerstone of community. It's exactly the reason I made sure Sun's covenant was not crippled in the same way.
I have now had several reports of Microsoft's representatives claiming their covenant is the same as Sun's; it is not, please make sure anyone who says so is challenged.
There's one more issue of note, which the NZ paper makes clear. Microsoft explicitly uses proprietary formats within their MS-OOXML specification (DrawingML for example). If they want to provide comfort to open source developers, they need to go further and cover all referenced formats with their "promise" as well.
Actually, yes - according to this Sun blogger Ubuntu already runs on it (see at the end).
Indeed, someone just did:
More details on Simply RISC's web site.
Well, apart from Simply RISC, who used the design to build a single-core chip (S1) for embedded applications.
And Polaris Micro in China, who are doing the same.
And David Miller & friends, who made Linux run on it.
And Canonical who support Ubuntu running on it.
And the other Linux distros picking it up.
And... Oh, sorry, you were just trolling, right?
The original T1 design was released under GPLv2 a while back, and the new T2 design will be released under GPLv2 for now as well. Using GPL v2 creates an implicit patent license so while using GPLv3 would make the situation cleared GPLv2 is probably sufficient.
And yes if you look at the map on the opensparc.net page (when they get some quota back after being slashdotted) you'll see they are getting a vast amount of interest from China, where I gather a company is already producing an OpenSPARC T1-derived chip for embedded use.
You're mistaking "committee members" with "supporters" I think. Neither IBM nor Sun was present to mindlessly support OOXML...
The plug-in includes some code that inserts the file format as an option in the "File Open/Save" dialog and in the rest of the file handling for MS Office so that the ODF formats appear as natural parts of Office (rather than in the un-natural way that the plug-in on SourceForge that MS sponsor tries to work). I gather that the developers are not sure if they can open that small piece of source code so the package as a whole is not under an open source license yet. The rest of the plug-in (the vast majority) is OpenOffice.org code and is open source.
Right - I've got a good handle on the license overall. However, there were some very significant changes in the last draft, final call and final version which may have significant consequences in our business (not least with how we are able to defend ourselves against patent aggressors) and we'll need to do some deep thinking about them.
Oh, we know where every line of code that's been Freed came from, have no fear! The engineering teams for both OpenSolaris and OpenJDK did a phenomenal job of source code archaeology. No, my point was that changing the file headers is actually a non-trivial task that triggers a wave of regression testing. Not to be undertaken lightly.
Yep, we were waiting for the final text. We engaged fully, contributed throughout but the final text was still news to us at the end of the week. And even when due diligence is done, we've still the need to get approval from the communities around the code in each case. I remain positive though, and I'll be trying hard to stay transparent (though some things just have to be kept internal of course).
You may be interested in the commentary I have written explaining Sun's position on GPLv3. Despite having been involved in the discussion process from the beginning, we still only saw the final text at the end of this week. Whereas some people (who ought to know better) think we can just change the licenses over-night, in fact we have to take into account:
So please do let Jonathan know how you feel, but be aware he's not the only guy who has to decide.
Actually, Sun has promised to release the Java platform under GPLv2, not GPLv3. I'll review GPLv3 when it comes out, though.
Sun will release the Java SE class libraries under GPLv2 (less the encumbered portions previously explained) before the end of June as promised.
In fact, the Java software is in Multiverse so has the same status as the JDK included in Debian's non-Free repository. Please don't ding Canonical this way. They have done more than you can possibly imagine to help Sun's staff with the work. They have worked faithfully with Sun's executives, persuading them of the need for Free software. They keep the pressure up by co-operating and setting the pace. They deserve your praise and not this criticism.
It's a bug, sorry - a placeholder installer was used during testing because of a licensing issue that needed sorting, and it didn't get replaced in time with the real .deb - we're working on it.
OK, I'm now feeding the troll, but I'll not let that smear stand. Sun has been scrupulously clear over exactly what is and is not under open source licenses. There are indeed parts of Solaris that are still encumbered, but that doesn't stop Nexenta and three other distributions building complete distributions, any more than the use of binary drivers by varying GNU/Linux distros invalidates the use of the term "Free" to describe them unless you are a complete radical - oh, sorry, it's me that's the zealot, I forgot.
The full source to the Java class libraries will be released in the spring under GNU/Classpath apart from the bits that hostile parties refuse to relicense to Sun. Fortunately the Classpath folks are helping work round that. Now, how about some working-with rather than all this negativity?
If you are actually interested and not trolling, I'd suggest you join the other 18,000 or so people over on OpenSolaris.org where the source code used to build Solaris is available under a Free license. You'll find a number of other OS distributions based on the same code - the most interesting is probably NextentaOS, which is essentially Debian with the kernel switched.
No, I got it, and I disagree with you. My amazement stems from the fact that some very smart people who I respect (the Linux kernel developers) are criticising the GPLv3 for things the discussion committees and the FSF are already addressing. I actually think that's clear from the original blog posting.
There's even more to it than that actually. First, the rights Sun has to Unix are far more extensive than those IBM has to AIX. In 1994, Sun bought an outright perpetual copyright license giving it non-exclusive ownership of the Unix source code. That's the basis on which Sun has been able to open source Unix in the OpenSolaris community. Sun has been fastidious in constantly acquiring all rights to Unix so that there could never be another party with a claim over Sun's core business.
However, that doesn't mean Sun (or any other Unix licensee) has had constant access to updates and new features. Whereas in the world of open source we all share enhancements with each other, in the world of Unix all the licensees retained their enhancements. The Unix copyright owner (originally AT&T, then Novell, then SCO) had the practice of buying enhancements (including drivers) from the licensees and then selling then to other licensees. This "clearing-house" role was what allowed Unix features to flow between the licensees without an open source community.
What Sun and all the rest of the licensees did was trade code with the Unix Licensor, usually for a net payment to the Licensor. That last payment was a confidential transaction like all the previous ones made by Sun, IBM, HP and others, but was a pretty routine transaction to secure updates and maintain a full license to the Solaris source. While I recognise that plenty of folks view that last transaction as "funding SCO", it was at its heart unexceptional apart from being the final rights acquisition before open sourcing the whole code-base. And having seen how SCO behaved since then, it was the last.
It's a hot topic and there are plenty of voices for and against. It was like this discussing the use of the GPL for the Java platform though, extensive and passionate debate right up to the last minute. In the end GPL v2 with the Classpath exception was clearly preferable, but every possible option was explored.
One key difference with Solaris though is that the base source code is already open source Free software and is in the care of the OpenSolaris community, so while Sun obviously gets a big say in what happens it's not just down to us, the community will also need to discuss it.
In fact my amazement stems from the fact that intelligent people who I respect are criticising things (like the DRM language) that have already been substantially altered, but arguing as if there has been no change to them. Sorry you find the whole thing "vacuous", you are clearly a Higher Intellect.
To see what I actually said, rather than relying on the ZDNet extracts, you'll find the original blog posting here. It's linked from the article but clearly from some of the comments above some people haven't spotted the link.
Seems to me that the main problem here is that there's no meaningful penalty for ignoring prior art. One idea that seems easy but I've never seen pursued is for the law to be changed to treat a failure to cite prior art as perjury. Then, should a successful prior art case be prosecuted against a patent, the applicant would be subject to fines or even imprisonment. This simple change would rebalance the system and result in far fewer lame patents with obvious prior art.
I suggest it would be wiser to infer that no package has so far been included in Debian with that license. There are plenty of licenses missing from that list.