The OSI License-Discuss mailing list has been ablaze for the past few days since Microsoft submitted its Permissive License to the OSI for official open source license approval. Jon Rosenberg, source program director for Microsoft, posted, “Microsoft believes that this license provides unique value to the open source community by delivering simplicity, brevity, and permissive terms combined with intellectual property protection.”
As much as I hate to say it, I have to agree that DiBona’s points are valid. None of what Hilf said impressed upon me that Microsoft was pursuing this as anything other than a marketing campaign. Attacking DiBona and Google certainly didn’t help matters much, either, especially given Microsoft’s track record regarding open source.
It’s a tough pickle, because the OSI is screwed either way. If they agree to review and approve the MSPL, Microsoft will feature it on its website alongside restrictive licenses and try to render them “open” by association, confusing your average CIO and getting a free ride on the back of a community they’ve constantly antagonized. Misleading (or helping to mislead) the public about open source and license legitimacy seems contrary to the goals of the OSI.
On the other hand, if the OSI turns MS away without even considering the license, they threaten to undermine their credibility and objectivity.
It will be interesting to watch this unfold.
To me it sounds more like “This is my ball and I decide who can play”
Definitely agreed. “And I if I don’t like your face, I will get my ball away and game is over.” 😉
The marketing excuse is ridicolous. If MS-PL is an open-source license, just state that, list and shut up.
So since when Microsoft is excommunicated from creating an open-source license? Blah.. usual oSS crap.
“If MS-PL is an open-source license, just state that, list and shut up.”
The whole point of the list is for them to discuss the proposal, in which, they are defining, is it an open-source licence. It’s funny how these things work, I know, it can tax the dumbest of people.
The whole point of the list is for them to discuss the proposal, in which, they are defining, is it an open-source licence. It’s funny how these things work, I know, it can tax the dumbest of people.
Reason why we’re buzzing about this is because OSI list is clearly revealing many people stating MS-PL should be rejected because… it’s by Microsoft. If it was discussion on legal-basis only, we wouldn’t be chatting about it. As MollyC stated, this would be a political reason, not based on merit.
And yes, it’s always wise to read the article we’re talking about before starting to type. I know not all of us could be wise… but…
How about, “license profileration”?
How about, “license profileration”?
That’s a different point. While I might agree (even if I think this is lost cause), there’s no point in stopping MS-PL while you have tens of others which are less than “rarely used”. Plus, you should then start to reconsider them all, if OSI really wants to be taken seriously. And in the list they’re not talking about “license proliferation”, rather about “stopping the enemy”-like. While Microsoft could live without its license approved, why should you turn OSI into a sidelined committee?
That’s funny, I read the article, and the newsgroup too!
Their discussion is never purely about legal matters.
How many people said it should be rejected because it is by Microsoft?
None? You’re right!
DiBona flagged some issues he wanted them to discuss, some OSI members suggested they not encourage that kinda thing, Hilf responded with “Google is nasty“ (yeah, I can paraphrase badly too
) and yet, still, no one said “reject it because Microsoft made it”.
You need to remember to not get mixed up in the political reasons.
That’s funny, I read the article, and the newsgroup too!
Their discussion is never purely about legal matters.
How many people said it should be rejected because it is by Microsoft?
None? You’re right!
Check better. DiBona went on with arguments like marketing, tactics and so on. And he’s been asked by other users why he was going to shift a technical (or legal) discussion into other kind of discussion which was not part of the matter. DiBona essentially wanted this license not be considered because it was from Microsoft, at least until MS stand about other things would have changed. Those things had nothing to do with merit of that license.
If you mean that he pissed off many people on that list, you’re right. If you mean that he didn’t want this license to be rejected (and even not considered) because it was from Microsoft, read better.
Yes, that’s the OSI’s job actually. But thanks for mocking it.
the bottom line is that this is ‘just another licence’. whatever Microsoft’s agenda may be, it’s up to the developers to use it or not. as long as the licence follows the criteria’s set by the OSI for approval submission then I see no problem.
developers have been perfectly capable to decide between open source licences sofar, I doubt that will suddenly change should Microsoft’s licence be approved. some people seem to think that developers just google up ‘licence’ when they’re ready to distribute their code and choose whatever pops up first. you really don’t licence your hard work on a whim.
… actually, when we were discussing licensing on our project, my boss really did do that — according to him, Google told him that the GPL v3 was the ‘best’ license.
People just choosing whatever license they’ve heard good things about without taking the time to read and consider the license is probably much more common than should be.
Almafeta wrote:
–“… actually, when we were discussing licensing on our project, my boss really did do that — according to him, Google told him that the GPL v3 was the ‘best’ license. “
nothing wrong in googling up licences. choosing a licence for your own hard work without reading it is another matter entirely, and I don’t think that is common at all.
not that you do either, it’s not hard to see where you are trying to go with this.
… I was going somewhere with it?
What “people”, Almafeta? Do you know _all_ people personally to make such generalizing statements? I don’t think so. And exactly what do you mean by “much more common”? How common is “common”? Can you give us some statistically verifiable data to support your claims?
Oh, people can come up with statistics to prove anything, Kent. 14% of people know that. – H. Simpson
the bottom line is that this is ‘just another licence’. whatever Microsoft’s agenda may be, it’s up to the developers to use it or not. as long as the licence follows the criteria’s set by the OSI for approval submission then I see no problem.
Well, for a start, the OSI were going to cut down on the shear number of licenses being submitted to the OSI by various companies, not just Microsoft, just to muddy the waters and get code flowing in their direction.
The OSI should be telling Microsoft that there are a great many licenses that already exist that are OSI approved, and they should just use one of them.
Microsoft has done nothing to help the OSI, or be supportive of their efforts in the past, so I fail to see why Microsoft should expect a free ride.
Edited 2007-08-18 23:47
Not quite, things are seldom so easy and transparent. The problem – as I see it – is mostly that MS already talks about being open source, and providing open source tools and apps, and whatnot, seemingly being deliberate in not differentiating between what licenses they use and what licenses are generally used in FOSS development. The reason for that is probably a PR and marketing one, that is trying to mix and blur the meaning open source among the MS customer base.
It doesn’t seem to be much, but if you take it as a defensive move, it’s not such a bad idea really.
I really don’t see an objective issue with refusing to listen to an almost (I’m being nice) criminal company from the perspective of OSS.
Chris is trying to choose the only winning move: “Not to play.” Microsoft is playing games and I believe OSI is small enough, young enough, and run by smart enough people to breaks its own rules and not play this hand.
Objectivity and unbiased opinions are for people who can’t learn and abstract, not to mention impossible goals.
Convicted of being a Monopoly is an admission the company is criminal.
However, they did have the biggest balls for over a decade by dangling their giant nutsack in the face of the public and getting away with it.
“Convicted of being a Monopoly is an admission the company is criminal.”
You might want to research the difference between “criminal” law and “civil” law before speaking on this topic again, lest you continue to look uneducated.
Oh, and “convicted of being a monopoly” is a meaningless phrase.
Edited 2007-08-19 04:54
The “informal” definition (New Oxford) of the adjective criminal:
“informal (of an action or situation) deplorable and shocking : he may never fulfill his potential, and that would be a criminal waste.”
He’s plenty close enough, regardless of the legalese.
“The “informal” definition (New Oxford) of the adjective criminal:
“informal (of an action or situation) deplorable and shocking : he may never fulfill his potential, and that would be a criminal waste.”
He’s plenty close enough, regardless of the legalese.”
I wouldn’t call bundling a browser with an OS “deplorable” or “shocking”.
Anyway, he’s not anywhere near close. He wasn’t speaking “informally”. He talked of “criminal conviction”.
Here’s one huge difference:
In order to be convicted of a criminal matter, one must be found unanimously guilty by a jury “beyond a reasonable doubt”.
In order to be found liable of civil infraction, one must only be found liable based on “preponderance of evidence” (i.e. 50% of the evidence plus one), a much lower burden of proof than “beyond reasonable doubt”. And such can be declared by a single judge, rather than a jury, who may or may not be a moron.
Here’s another difference. Microsoft was found liable for actions taken *before* it was ruled that they had a monopoly in a particular market. And then they were punished for those actions in a retroactive fashion. Such is not possible under criminal law. (This is one of the flaws I find in antitrust law; due to its retroactive nature, you have to second-guess every action you take at a given time for fear that some judge *might* declare at some point in the future that you had a monopoly at some time in the past. It’s BS.)
And no, this isn’t simply an academic debate.
The reason that antitrust law is civil law and not criminal law is that antitrust law has nothing to do with good or evil. It’s not inherently evil to do anything that Microsoft did. It was only that since they had a dominant position in a particular market, that those actions were deemed harmful for the marketplace. Not “evil”, just harmful. Using words like “criminal” carries the “evil/deplorable” baggage, which is inappropriate for this case.
Anyway, it’s not relevant to the OSI submission process, despite DiBona’s wishes to the contrary.
Edited 2007-08-19 16:10
Anyway, it’s not relevant to the OSI submission process, despite DiBona’s wishes to the contrary.
I’m afraid it is. A track record with an open style organisation such as the OSI is pretty much a given. Microsoft has done nothing to help the OSI, has done everything to deride it’s cause and that of the concept of open source software and the licenses they are submitting do nothing to help open source software, nor do they further it in any way. The OSI has also stated it wants to cut down on the variety of different licenses.
I’m afraid Microsoft cannot just submit a license to the OSI and expect it to be accepted. The licenses many vendors have come under quite a bit of scrutiny as well, and arguably, they shouldn’t have been approved by the OSI.
Anyway, it’s not relevant to the OSI submission process, despite DiBona’s wishes to the contrary.
I’m afraid it is. [snipped trite Microsoft bashing]
If you bothered to read the OSI discussion list, you’ll see that DiBona isn’t getting much, if any, support. So while you believe this is relevant, I don’t think OSI on the whole does.
I’m afraid it is. [snipped trite Microsoft bashing]
Whatever sweetheart.
If you bothered to read the OSI discussion list, you’ll see that DiBona isn’t getting much, if any, support. So while you believe this is relevant, I don’t think OSI on the whole does.
Di Bona has got quite a bit of support once he’s explained his position, so please don’t take that Microsoft tac with me and tell me what’s going on as if it’s some sort of fact.
Fact is, that view is very important to the OSI and the licenses it endorses, and having a track record is very important in a community and committee based organisation. If you don’t have it then you don’t get a free ride. I also explained the OSI’s current licensing view of cutting down on the number of needless licenses, most variations on a theme, and that really does apply.
I’m sorry you believe criminal law has to do with good and evil. It doesn’t. Personally, I consider greed and monopoly action to be immoral (evil for me), but that wouldn’t make it criminal even if everyone agreed with me.
That said, he was speaking informally; welcome to an Internet bulletin board. I suppose there was some formal context in the point, and in order to consider it as a real reason for the OSI to do something it would need to be formal. But I don’t think he was trying to be formal, and neither way I (not about that at least).
We were simply trying to show just how dangerous this situation may be. But you’re right, this is hardly an academic discussion. It’s really an argument between those who respect Microsoft as a dangerous entity (fear) and those who value objectivity above that. It’s unlikely to reach a compromise the parties can agree on as well, but then again what bulletin board argument ever has
.
“I’m sorry you believe criminal law has to do with good and evil. It doesn’t. Personally, I consider greed and monopoly action to be immoral (evil for me), but that wouldn’t make it criminal even if everyone agreed with me. “
Maybe my comment was too general. I don’t believe that criminal violations are the only evil acts. But nor do I believe that all illegalities are immoral.
In particular, antitrust law is not about morality, it’s about trying to maintain a vibrant marketplace, for the purpose of benefiting consumers (not competitors), period. That’s it. Nobody is going to “hell” for bundling a browser with an OS. (Nobody is going to jail for it either, for that matter.) I can’t think of any of the 10 Commandments that such an action violates (and I can list them, unlike those that want them taught in schools and displayed in court houses
). It *might* harm the marketplace, but it’s not a “sin”.
Now, I don’t know what you mean by “monopoly action”, but “greed” is indeed one of the seven deadly sins. But name me a corporation that isn’t greedy. I hate to quote Gorden Gecko, but “greed is good”.
The entire economy is based on it. It’s why capitalism succeeds and communism fails. Charity and philanthropy are clearly virtues worth having, but greed is required to survive. Everyone has both greed and charity in him. And Microsoft is no more “greedy” than any of its competitors are (though they may be more charitable).
When I think of an “evil” corporation, I think of the likes of Enron, IG Farben, WorldCom, polluters, child labor exploiters, etc. Those are examples of truly “evil” corporations, not Microsoft.
Now, I could raise questions as to whether a company making billions on the backs of a free labor force is “evil” or not, but I won’t go there (right now, anyway).
Edited 2007-08-21 06:42
“I really don’t see an objective issue with refusing to listen to an almost (I’m being nice) criminal company from the perspective of OSS.”
You’re not being nice, you’re being discriminatory. IF MS follows the process and the rules, their past behavior is somewhat irrelevant. OSI’s job is to verify and certify Open Source licenses. They just can’t reject it out of hand, that shows they are not objective, and it undermines efforts to get everyone, including MS to play by the rules. Getting MS to play fair with OSS is a goal I can get behind. That way everyone wins.
“Objectivity and unbiased opinions are for people who can’t learn and abstract, not to mention impossible goals.”
That’s pretty closed minded, and a bit retarded as well. For an organization like OSI, they have to be definition of objectivity, or they are useless. These issues are not abstract, they affect the way thousands of people and organizations do business everyday.
The license and the license’ owner should both be under investigation. If Microsoft is going to happily give up all ownership and claims to ownership (including trademarks) then we can evaluate just the license.
Sure, no problem, the license should be evaluated and passed if it meets OSI’s criteria, but it should also not be denied because of who made it, if that is the only reason
Objectivity and unbiased opinions are for people who can’t learn and abstract, not to mention impossible goals.
By that logic, people who are able to learn should always form biased opinions. Except that the definition of “biased” means being unable to learn to see things in new ways as new situations present themselves.
Sorry, I just don’t buy your twisted argument. Microsoft’s past behavior should not be the basis of refusing to review their license request. As other people have noted, just letting them create an open-source license does not immediately mean that thousands of developers are going to flock to it. Now I’m not saying they should approve this license or not, as I myself have not reviewed, it but I am saying that it should at least be looked at carefully before being rejected.
If OSI were to reject the MS licenses for political reasons, then they would reveal themselves as a political organization and lose all credibility. If they reject the licenses they better have sound reason for it. From reading the OSI discussion list for both MS-PL and MS-CL, it looks like both are on the way to being approved.
And there was no “attack” on DiBona or Google. DiBona decided to inject politics into the debate, so he left himself open to likewise response. Regarding Google’s use of open source licenses, one issue (I don’t know if this is what Hilf was referring to) has to do the fact that Google uses tons of GPL in its products, releases those products to the public for use, but doesn’t release their own code because the proucts are web apps rather than local binaries, which one could argue goes against the spirit (if not the letter) of GPL.
Edited 2007-08-18 22:15
“If they reject the licenses they better have sound reason for it.”
Oh, you mean, like, a bunch of already existing OSI-approved licenses with the same features that can be used by Microsoft’s license? Funny how Microsoft insists on having their own little license, don’t you think? Almost, like, they’re trying to be… oh, I don’t know… political when it comes to licensing issues.
“DiBona decided to inject politics into the debate…”
What “politics”, Molly? What politics has to do with the facts that to this day Microsoft continues to spread misinformation about the nature of open source software, especially that licensed under the GPL (an OSI-certified license, by the way), even as they seek OSI’s approval of their license?
This is a company well-known for its opposition to all things open source, and now all of a sudden they want to be approved by the very entity they’ve spent years badmouthing. Interesting.
“Oh, you mean, like, a bunch of already existing OSI-approved licenses with the same features that can be used by Microsoft’s license? Funny how Microsoft insists on having their own little license, don’t you think? Almost, like, they’re trying to be… oh, I don’t know… political when it comes to licensing issues. “
From reading the discussion list, I don’t think that will be an issue. For one thing, MS’s license include patent coverage.
Personally, I think the “non-proliferation of licenses” thing is bull. If a party feels that existing licenses do not meet its needs for whatever reason, and creates its own license that meets all of OSD guidelines, and submits it for approval, then OSI should approve it or they should prove to the party that some other license fits their needs. The OSI doesn’t know all of the needs and nuances that a party might want, so it would be difficult for them to tell someone that some already approved license fits their needs.
But even putting that aside, if the party submitting the license can come up with even one scenario for which other licenses either do not suffice, or do not deal with the scenario as well as the proposed license does, then OSI should not reject the license on the basis of non-proliferation.
Anyway, many of the existing licenses aren’t in use at all, while MS’s are already in use. If OSI is concerned about license proliferation, they should reject one of those old licenses that nobody is using to make room for new licenses that are in use.
But let’s get down to brass tacks. MS haters, such as yourself, can’t find any real reasons to reject these licenses, so you’re grabbing the “non-proliferation” thingy for dear life, so desperate you are to find any reason at all to reject these licenses. That is quite clear. But unfortunately for you, I think this is such a high-profile case, that OSI wants to approve or to reject the licenses based on the licenses themselves, not based on politics, and not based on some declaration that other licenses suffice. Even in the mailing list, there is debate as to what the non-proliferation thing actually means, and it appears that no license has ever been rejected on that basis. Suddenly doing it for MS’s licenses would look mighty suspicious (and a bit too convenient for MS-haters).
Also note that OSI members themselves were calling for MS to submit MS-PL earlier this year, so it would be problematic for them to now say that other licenses suffice for the needs that MS’s address.
“This is a company well-known for its opposition to all things open source, and now all of a sudden they want to be approved by the very entity they’ve spent years badmouthing. Interesting.”
Huh, I thought you guys wanted MS to start moving to OSS. Guess that was a bunch of BS. MS-haters moving the goalposts, as usual. (BTW, your statement that Microsoft is against all things open source is propaganda, nothing more; or how do you explain the many projects MS has already released the code for, including important things such as the DLR?)
What are you going to do when these licenses get approved, cry?
Edited 2007-08-19 06:55
I might be missing something, and I hope someone will read this post buried so deep down on the discussion and clarify this: in practical terms, how is the MS-PL different from the CDDL, which is already OSI-approved? I’m truly surprised that so far NO ONE brought up the obvious similarities to the CDDL!
1) License covering the code in question but allowing mixing with other licenses: both check
2) Patent coverage: both check
3) Copyleft: both check
4) Unambiguous language: CDDL has the edge
5) Shorter expression: MS-PL has the edge, but this isn’t such a crucial advantage
5.1) Shorter than the GPL: both check (yeah, I included this point tongue-in-cheek-ly
)
6) Recognition and broader understanding by the community: CDDL has the edge, as it’s based on the very well known MPL and extended not to be explicit about Mozilla
7) Regarding new versions of the license: CDDL is extremely explicit about this
8) Amount of code already covered by one of those licenses (call it the “battle- & time-tested” point): CDDL clearly has the edge, by a wide margin
So, again, what is it that MS-PL brings to the table? Being a handful of paragraphs short? While still carrying language many (read the comments to this very news post, as the ones on the OSI list and the Linux.com article) considered ambiguous?
The OSI doesn’t know all of the needs and nuances that a party might want, so it would be difficult for them to tell someone that some already approved license fits their needs.
Well, if the OSI reviews the terms of the new, requested license, they should get a pretty good idea of whether an existing license meets those needs, shouldn’t they?
If there are some other licenses that look very nearly identical to the MS-proposed one, then any nuances between it the MS one should be apparent based on MS’s license description. But the idea of MS creating a license that looks just like an existing one, but where the important nuances are simply left out only to be “found out about later”, or are only known about by the highest-ups of Microsoft’s legal team, is intolerable.
Considering that the license proliferation project seems to have been active in OSI since 2006, I don’t really think it is “bull”, or has anything to do with MS specifically, especially their recent submission. Let’s not discount its importance just because you think someone is an MS-hater.
The license proliferation thing has no hard and fast rules; it’s like ad-hoc law. And as such, license proliferation must be weighed on a case by case basis. For a particular case, do the benefits of approving a license outweigh the harm of proliferation (assuming that the licenses under consideration are no different than what has already been approved, which I don’t think is the case here anyway).
BTW, anyone with half a brain can see that the license proliferation objecting is being raised only because anti-MS folks can’t find any real objection to these licenses.
Edited 2007-08-20 19:39
Actually, I think the whole license proliferation thing was a result of the OSI folks saying, gee, “I think we messed up in the past by allowing all these vanity licenses”. And the reason there are no hard and fast rules is because they are still working on the project. These licenses just came up at a time when they hadn’t come up with those rules yet. And you can’t call everyone who wants to talk about it an MS hater…you MS apologist you..
Funny how Microsoft insists on having their own little license
I am not sure that there are many OSI licenses that grants patent and copyright protection. MSPL is, at least uncommon, if not unique. That justifies its existence.
What politics has to do with the facts that to this day Microsoft continues to spread misinformation about the nature of open source software, especially that licensed under the GPL
That is politics.
What Microsofts, or anyone elses preferences about OSS have to do with license submittion process ? If there are regulations, they should apply equally to everyone, regardless of who they like or like not. There are no “good and evil” here.
“Oh, you mean, like, a bunch of already existing OSI-approved licenses with the same features that can be used by Microsoft’s license? Funny how Microsoft insists on having their own little license, don’t you think? Almost, like, they’re trying to be… oh, I don’t know… political when it comes to licensing issues.”
Sun has it’s own license, CDDL. Apple has it’s own license, MIT has their own license. X11 has it’s own, 3 versions now of the GPL, there’s the LGPL, Apache has their own license. There may be too many licenses, but many organizations have wrote their own, and had them passed. Not passing MS’s license for no good reason other than, there are existing licenses, go use one of those is just political nonsense.
Companies like to define things in their own way, it gives them that feeling of control that companies love. If it gets MS to start playing ball, let them have their license. It doesn’t hurt anybody.
If OSI were to reject the MS licenses for political reasons, then they would reveal themselves as a political organization and lose all credibility.
Before you go crucifying the OSI for something you don’t even know happened, I have to remind you that microsoft is a software company turned political organization (see link).
http://www.osnews.com/permalink.php?news_id=18098&comment_id=248910
Why the double standard MollyC?
“Before you go crucifying the OSI for something you don’t even know happened, I have to remind you that microsoft is a software company turned political organization (see link).
http://www.osnews.com/permalink.php?news_id=18098&comment_id=24…
Why the double standard MollyC?”
This isn’t a double standard. If OSI wishes to approve or disapprove based upon the views of one of Microsoft’s competitors, do you even realize how OSI would have no credibility anymore? Do you realize that OSI would just be seen as a mouth piece for Google? Can you even fathom the PR gift this would be to Microsoft?
What is a double standard is how so many people cry and whine about freedom of choice, yet really we see honestly that their version of freedom means to disallow anything Microsoft….and this makes you different than what you accuse them of being how?
Reading the OSI License-Discuss mailing list archive I discovered discussions about GPLv3.
If i understand it correctly the GPLv3 was submitted by Chris DiBona. I’m a little bit confused since if i remember correctly OSI only starts their approving procedure if the license in question was submitted by the license author.
It seems like FSF hasn’t submitted the GPLv3 to OSI so why does they start their approval procedure for it?
IIRC, FSF doesn’t agree with the OSI’s existence, because they believe Open Source != Free Software.
>IIRC, FSF doesn’t agree with the OSI’s existence, because they believe Open Source != Free Software.
This wasn’t my question!
My question is:
If OSI only start their approval process for licenses submitted by the license author than why do they start their approval process for GPLv3 even if it wasn’t submitted by the author (FSF)?
I can’t answer because … http://opensource.org/approval does not mention the process you are talking about, so I assumed you’re ill-informed.
…But this is what a proper Open Source License needs to look like. Regardless of Microsoft’s intentions, I think people shouldn’t try to game Microsoft on this one, even if they are trying to game the system. A License that fits on one page (in my browser). The holy grail indeed.
You may be right. But I, like DiBona, just hate to see OSI being used to enable its most vicious competitor.
I agree with you 100%. Microsoft’s intentions are really irrelevant in this matter. The focus should be on the merits of the license itself and whether or not it meets the standards of OSI. It shouldn’t matter whether the license was submitted by Microsoft, you, me, or somebody from a small town in the middle of no-where Kansas. Saying, “This license is not really valid because it was submitted by Microsoft” is in and of itself not really a valid statement and shows a decision based off of nothing but emotion.
The MS-PL is clearly compliant with the OSI terms. It should be approved.
MS-PL will most likely only see limited usage because of its incompatability with the dominate GPL and LGPL free licenses and many other “copyleft” OS licenses. The name of the license itself is a deterent to all projects not closely associated with Microsoft products and/or platforms.
Except for Microsoft employees, it really is the last place developers, who are interested in Open Source, will turn to for a license selection and advice. Lets just say Microsoft’s credability in the OS movement is insufficent at best.
Edited 2007-08-18 21:40
If, as the article says, the MS-PL licence is very similar to a previously-approved licence, why can Microsoft not use that licence with an exemption clause?
From a business perspective, it’s usually best to have had a say in the writing of a license you’re using.
From a legal persepective, when you add on extra conditions to a pre-existing license, you’re adding a layer of confusion that could lead to trouble.
Additionally, ‘BSD license’ doesn’t describe just one sort of license, but a number of closely related licenses, since nobody owns a copyright to the ‘BSD license’ (despite its being associated with a commercial term, a ‘BSD license’ is more of a concept of a certain kind of permissive license than it is a license in and of itself). Calling the MS-PL license a BSD license, while it would be correct in spirit in that it’s about as free as free software can get, would not be technically correct, since the MS-PL’s patentlefting clause adds in a layer normal BSD licenses don’t have.
Besides, there are several different sets of expectations when people say ‘BSD,’ which might spread confusion if they linked the two in people’s minds — and when it comes to legal issues, Microsoft runs a very tight ship (which is probably why they’re not marketing the MS-PL as a BSD-like license).
That said, I believe MS has used the BSD license in the past…
That said, I believe MS has used the BSD license in the past…
AFAIK, they have, and even complied with requirements to credit the authors where needed when that version of the BSD license has been used.
Ok, well, that’s interesting and all, to the extent which a license can be without being associated with any content.
Does anyone know if MS offer any actual code under PL?
Microsoft has lots of MS-PL projects.
According to Bill Hilf’s post to the cited discussion list http://www.nabble.com/RE%3A-For-Approval%3A-Microsoft-Permi… :
“There are already several hundred community projects that use these licenses, including over 150 Microsoft projects.”
A very prominent recent example being the DLR (Dynamic Language Runtime), that allowed Moonlight (Mono’s implementation of Silverlight) to be created so quickly. Also, IronPython and IronRuby code are released under MS-PL.
Lot’s of MS-PL and MS-CL projects from both MS and 3rd parties can be found here:
http://www.codeplex.com/
IronPython is on CodePlex (http://www.codeplex.com/IronPython), but IronRuby will be put on RubyForge in the next few weeks.
MS also has some projects on SourceForge.
Edited 2007-08-18 22:04
Heh.
Just look at the number of GPL projects on CodePlex.
Well at least Microsoft was nice enough to allow license selection.
Mhmm, nearly 12% of their hosted projects.
“Mhmm, nearly 12% of their hosted projects.”
How are you guys determining the number of GPL projects on CodePlex? I must be blind, but I can’t find anything on there that searches by license.
Edited 2007-08-19 00:56
Googling and simple mathematics.
See my other post.
Where did you get that number?
I dont disagree that it might be accurate but I couldnt find it myself.
Anyway, there are 82,700 GPL/LGPL combined projects on SourceForge, some being mixed with other licenses.
You’re right, it is a drop in the ocean compared to other project sites.
1070 results from Project search
122 results from GPL search
and some heavy rounding
’course, my numbers are prone to error. But isn’t all marketing.
Well that number doesn’t sound right to me because there are 2,053 projects on CodePlex and searching for GPL gives me 5 projects. But I’ve scanned just a few projects yet saw more than five GPL licensed projects and then a couple BSD, and MIT licensed projects. So the search mechanism is clearly ineffective when querying for projects by license.
Edited 2007-08-19 16:56
nb, I didn’t use the codeplex search.
I was hoping to find some basis for those numbers on CodePlex, but it’s a little tricky, MS hate citing their sources.
(site:codeplex.com/*/Project/ “Microsoft Permissive License”)
Gives 338 hits in Google. Do we subtract the 150 claimed by Microsoft? or do we take that as an additional 150 ontop of our 338 figure?
(site:codeplex.com/*/Project/ProjectPeople.aspx Microsoft)
Gives 1070 hits in Google.
Ah, a problem, site copyright notice. But at least we have a total for projects.
Looking at the Microsoft user profile gives us ~180 projects they coordinate on the site.
So I guess their figures are psuedo-real. Assuming MS pick MS-PL for all their works, 338 – 180 = 158 projects using MS-PL. Maybe that is several hundred to them, but meh.
There are very good and big projects released under MS-PL. I know because I use some of them. We’re not talking about silly nift libraries: we’re talking about complete CMS systems, languages and stuff. Can’t say if they are several hundreds. I can say there is tens of thousands of $$ in work there.
I’ve also stumbled into many projects not coordinated by Microsoft which are using MS-PL.
Are you assuming that every MS-PL project is hosted on CodePlex?
You’re welcome to provide numbers outside of the scope of codeplex.com, but as it stands:
Sourceforge has 1 project.
Berlios.de has 0 projects.
Ohloh.net has 0 projects.
After that, I thought it pointless to go find individual sites for additional +1s.
I guarantee that if it passed OSI, maybe even if it doesn’t, they will. They’re not stupid. They’ll at least offer some projects they don’t believe will sell. Good ideas might be visual studio (which they only sell now in order to not anger customers who already paid, and to keep Borland fro suing) or MSN messenger.
But I doubt you’ll see anything that they’re charging for now or believe they could charge for later. Keep in mind, I’m not sore about this, it’s just part of reality for Microsoft: They have a business model to follow.
Does anyone know if MS offer any actual code under PL?
IronPython is under the Ms-PL:
http://www.codeplex.com/Wiki/View.aspx?ProjectName=IronPython
From IronPython’s site:
IronPython is a new implementation of the Python programming language running on .NET.
So it is a re-implementation of a Free programming language running on a non-Free platform? gr8. Does it run on Mono as well?
Yes. The Mono project regularly refines its infrastructure so that the major .NET languages will run on Mono. IronPython is one of the main languages they test compatibility with. The IronPython developer community has assisted with this.
So, IronPython is a Free reimplementation of a Free programming language that runs nicely on a Free platform. And it’s from Microsoft. Live with that.
But thing is, those hastily repeated “Free”s in that sentence don’t share the same meaning.
Yes. And it’s faster than CPython. Unfortunately it doesn’t have all of the native python packages that were written in C as far as I know. But this is not the only thing they have opened up on there.
Don’t expect the Windows Kernel to appear on CodePlex anytime soon, but don’t mock a small team’s hard work either. Unless you have invented your own free programming language or can even demonstrate knowledge of how much work is necessary to implement one, you just look like a greedy freeloader.
don’t mock a small team’s hard work
True.
I appreciate these kind of efforts as much as the next guy, but I am still a little perplexed about the decision to re-implement the language and duplicate efforts instead of improving CPython for the benefit of an even broader audience. However, I also realise there might have been perfectly good reasons that are not very obvious for a by-stander like me. Maybe C# makes it possible to maintain a smaller but equally functional code-base than C or brings higher performance in particular cases, maybe CPython did not suit theirs or others particular needs in terms of Windows platform integration? The benefits seem to outweigh the consts. I don’t know, I use neither CPython, IronPython, .Net or Mono for my cross-platform programming and I don’t really care about Windows. What does IronPython do that CPython does not?
I can only hope that being a Free cross-platform python application programmer has become easier and more interesting with their efforts, because this I do care about, and if so, I congratulate them. Judging from your user name, I guess you do to 🙂
I find it a little odd though that IronPython started out as GPL and ended up with the Microsoft-PL and hope the implementation is not part of a embrace-and-extend strategy. Maybe someone with specific insight into these matters could elaborate a little.
Unless you have invented your own free programming language
It seems to me that they didn’t either
or can even demonstrate knowledge of how much work is necessary to implement one
Dude, you are seriously expecting me to be able to convey that in a post on OSsews? I can only say that I know enough programming to get a vague idea of the gargantuan efforts involved, but too little I guess to easily see that their efforts were really necessary in the first place.
you just look like a greedy freeloader.
C’mon. Cool down. A freeloader is someone who uses something repeatedly without contributing even though he or she could. I can’t see how I qualify as one in this case.
I may be grumpy, but not a freeloader.
“What does IronPython do that CPython does not? “
IronPython runs on .Net and Mono, just as Jython runs on JVM (both originally written by Jim Hugunin). CPython does neither of those things. And as such, IronPython code can interact with code written by any other .NET language, and can call all .NET framework libs. IronPython also allows Python source code to be directly used by Silverlight web pages.
Here’s a history of IronPython (this regards IronPython 1.0. IronPython 2.0 is being developed right now, the main difference being that it runs on the DLR):
http://blogs.msdn.com/hugunin/archive/2006/09/05/741605.aspx
Here’s a screencast of IronPython working directly with C# and VB:
http://weblog.infoworld.com/udell/2006/08/30.html
And here’s a demo of IronPython 2.0 working with IronRuby, JScript, and VB, all with integrated source code debugging using each language
http://sessions.visitmix.com/upperlayer.asp?event=1011&session=2012…
Anyway, are we really to the point where every lanuage must have only one implementation now? So there should only be one C compiler, one Pascal compiler, one Lisp, etc? I hope not.
Edited 2007-08-19 10:00
Yes, IronPython runs on Mono very well, and it did so for more than a year now.
Details are here:
http://fepy.sourceforge.net/doc/ironpython-mono-report.html
The OSI license review debate surrounding Microsoft’s submissions has focused on the scope and criteria of the review. Is OSD-compliance sufficient or are issues like license proliferation and certain vendors’ past indiscretions important to consider as well?
It’s important that being an OSI-approved license means something specific. They can’t go around making exceptions for or against certain vendors. They can’t continually raise the bar for new submissions to counteract license proliferation. If OSI wants to become more rigorous in their criteria, then they need to apply their new standards to previously approved licenses.
However, it’s not clear that OSI actually follows their Open Source Definition (OSD). I specifically have problems with section 2, which requires the availability of corresponding source code. Several OSI-approved licenses don’t satisfy this criteria, including most “permissive” licenses. For example, the BSD, Apache, and Ms-PL have no clauses that require the availability of source code corresponding to covered works distributed in binary form.
Even the GPL appears to fall short of OSD section 2. There several prominent instances where GPL source code is deliberately obfuscated. The nv graphics driver is a well-known example. The GPL doesn’t explicitly prohibit obfuscation. It only requires that the source code be machine-readable and in the preferred form for modification. The OSD has this same exact requirement, but goes further to prohibit obfuscation.
But corresponding source code is a dark corner for the FSF as well. While the availability of source code is explicitly listed as a precondition for freedom 1, the FSF considers permissive licenses to be free software licenses. It’s highly questionable whether you can be free if your software is available in binary form only.
It’s certainly questionable whether software can be open source if the source code is unavailable. Most software under the BSD, Apache, and similar permissive licenses is available in source form, and thus complies with both definitions. But the licenses themselves clearly allow for software which is neither open source nor free.
<blockquote>
Even the GPL appears to fall short of OSD section 2. There several prominent instances where GPL source code is deliberately obfuscated.
</blockquote>
What are you going on about? Obfuscated is not the ‘preferred form’ for modification. Otherwise, people would distribute assembly language or even machine code, as these are merely obfuscations of the source code.
Just because someone claims to satisfy the GPL by distributing obfuscated source code doesn’t make it so.
The “preferred form” was intended to refer to the way in which the source code is delivered. In other words, a tarball of text files is a preferred form. A ream of paper or an e-book is not.
Projects do distribute assembly code as source code (machine code would be considered object code), and it is not necessarily obfuscated. Often times it is the only way to write the software. Distributing assembly merely as an obfuscation would go against the OSD.
Perhaps a case could be made against GPL projects like the nv driver by stretching the interpretation of “preferred form”. However, until it is clear that the GPL prevents deliberate obfuscation in practice, it cannot be said to comply with the OSD.
The nv driver is not actually under the GPL, but an MIT license.
Edited 2007-08-19 00:11
I completely agree about the Open Source licenses. The new BSD *license* is not Open Source (meaning it doesn’t require distribution of source code), even if all projects using this license are kindly providing the source code.
Regarding the MS-PL, it’s NOT a BSD-like license. It’s a *copyleft* license. I’d say it’s a (bad) LGPL-like license that doesn’t require distribution of source code.
Is it a Free license? I’d say yes, considering all the other Free licenses out there. It’s not different from them in any fundamental way.
Is it a good license? NO, it’s a bad license. First because we don’t need another copyleft license in the FLOSS. Second because it doesn’t require source distribution. Third because it’s confusing in two ways:
– First by MS comparing it with the BSD ans Apache licenses.
– Then by the license itself “hiding” its copyleft nature by not explaining it well enough (and yes, any derivative work must be distributed under this same license, in case anyone has a doubt).
Certainly a bad license. Knowing who wrote it, I’d even say it’s an evil one (and written with that very purpose).
This is a balanced comment, even if I don’t agree on all points.
“Hiding its copyleft nature” is one of them. This license is very simple, it’s not hiding anything. You would say that BSD license is hiding its copyleft nature as well?
(and yes, any derivative work must be distributed under this same license, in case anyone has a doubt).
I’m not sure about this. MS-PL states that *if* you distribute source code, you must distribute it under MS-PL itself, while if you distribute in binary/compiled form, you must distribute under a license complying with MS-PL. Complying mostly means retaining copyright notices, no trademarks, patent litigation limitations and as-is condition. That means you can mostly distribute as you wish if you just include the patent-litigation limitation and the no-trademarks ones.
I still can’t see why people is buzzing about this license and someone also say it must be rejected. It’s clearly an OS license and must be approved. Then, if we want to start the reduce number of approved licenses game, we have to start re-examining all approved licenses. And I bet MS-PL won’t be the most unuseful among them…
This license is very simple, it’s not hiding anything.
It’s simple, but confusing. First, the MS guy who presented it said it’s a BSD license, which is not true. Then it’s also very unclear about your modifications, should they be under this same license? Nothing explained, so it looks like it’s not copyleft when it really is (IMHO). This is what I meant by “hiding” its nature.
Complying mostly means retaining copyright notices, no trademarks, patent litigation limitations and as-is condition.
You forget to comply with the condition that talks about the conditions for redistribution. If you want to redistribute your binaries under a license that *fully* complies with this one, you must include that too. What does that mean? That you really should distribute your binaries under this very same license. Not strange, any copyleft license *only* complies with itself. So you see, again trying to hide the truth.
So you see, it’s very simple but very misleading. And since it’s not written by amateurs I’d say it’s done on purpose (and hence my comment about it’s possible evil nature).
You forget to comply with the condition that talks about the conditions for redistribution. If you want to redistribute your binaries under a license that *fully* complies with this one, you must include that too. What does that mean? That you really should distribute your binaries under this very same license. Not strange, any copyleft license *only* complies with itself. So you see, again trying to hide the truth.
I’m not sure about this. A license complying with MS-PL means something which retains its basic principles. This means you cannot allow anyone to use trademarks of original developer, for example. But you can allow them to use yours.
Or that you can’t allow thirdy-party to keep using original software part of your derived work when they start a patent litigation over original software and so on. But could you allow thirdy-party to keep using this software if they start a patent litigation about your part of a derived work (i.e. the code YOU added before re-distributing)? I guess you can.
I’m not a lawyer but it looks to me that they set specific requirements which you need to mantain over original code. These aren’t very intrusive (as for GPL v3, for example) as most of them are simply aiming to provide safe dealing of original software relating to patents, software AS-IS and trademarks and copyrights. So basically, except for these, you’re free to do anything with original code.
I’m having a little trouble finding the part of the MS-PL that makes it a copyleft license. I can see that about the MS-CL, but not the MS-PL. Can you quote the relevant section and explain your reasoning, or provide a link to such an argument made by someone else?
I’m having a little trouble finding the part of the MS-PL that makes it a copyleft license.
Say you get some BSD licensed code to create a derivative work. You want to distribute your derivative work under the GPL. Can you? Yes, you can, because the GPL is copyleft, but BSD isn’t, so they are compatible.
Now imagine the original code you want to use is MS-PL. Can you distribute your derivative work under the GPL. No, you can’t. Because both GPL and MS-PL are copyleft licenses.
I hope that helps.
OSI has to agree to the license if it meets the criteria or MS has won?
Well not really what I wanted to discuss but thought I’d let it hang in the air.
If MS gets these licenses approved, it does not matter for the anti MS people nor for enterprises. Why? The CIO has either made up his mind or will ask a geek that knows Linux within his IT dep. The CIO might ask some analysts org, like Forrester or others, they would be able to filter out the noise and give some advice+ tell the CIO that they should either have a lawyer look at the licenses to “adopt” or agree on only certain licenses to be used and what the risks are with some licenses.
If the MS license is a joke, can’t really use it nothing to fear. If it’s fine, let’s discuss something else 🙂
A tip: ask your CIO/CTO if they even know about MS open-source efforts…
Let’s assume MS want to use oss when they have to because they see there is no way back, they can’t flag the patents.. suing RedHat means Oracle,HP,IBM etc etc..
MS will start to open up when they need to to close a business deal and nothing more, nothing to see, nothing to discuss. Nice to see OSNEWS having this advert:
State government says Linux was too big a risk… right below “submit comments” this is PR/Marketing because MS is loosing business to OSS and the licenses are nothing more.
Google is perfectly exploiting the open source community’s MS hate. My question to google is, if you are that friendly to open source then:
1. Where is the modified linux kernel that you use on your servers?
2. Where is the source code for your search engine?
3. Where is the source code for other applications you build like google desktop etc.
Sorry google is not a company I would like to defend for pro open source stance.
“1. Where is the modified linux kernel that you use on your servers?”
Note that GPLv2 under which linux kernel is licensed allows making private modifications without releasing them.
However, Google also sells “Google Search Appliance” which runs their modified linux kernel. Still, GPLv2 only requires them to distribute the source to GSA customers. But Google is being nice, and you can download all modifications to open source softwares on GSA here:
http://code.google.com/mirror/gsa.html
I hope that answers your question adequately.
No it really does not answer my questions. You have avoided all the questions actually.
GSA is not the same kernel that Google runs on it’s servers. Go check with google employees.
Google has nothing to do with this thread outside of the salary they pay Chris DiBona. Please, get back on topic.
The MS-PL actually isn’t a terrible license.
The key point it has over the X11 license [note: “BSD” is ambiguous, as is “MIT”. When you say either, you probably mean the X11 license—you’re better off just saying that if you do to lessen the potential for confusion] is the patent grant.
“If you distribute any portion of the software in compiled or object code form, you may only do so under a license that complies with this license.” is quite odd wording, though I know what it means. It actually reads as though your binary license itself has to be distributed under the terms of the MS-PL, but that obviously isn’t case.
Other than that, though, there’s not a huge amount wrong with it. It’s not like it’s one of the really crappy MS licenses, after all.
My take? If it meets the OSD, approve it. If Microsoft start using “OSI approved” as a selling point for some work it does (provided it doesn’t misrepresent anything), then that’s a good thing, not least in the eyes of the CIOs that the OSI have been trying to target forever, with varying degrees of success.
If the OSI want to get into “past practices” and such, IBM and Sun have been no angels either, no matter how much they’ve contributed to the open source & free software world; it’s probably a bit late for that. They could, however, make future conduct a conditional of ongoing approval—and apply that to everybody—but to be honest I’m not entirely sure that the OSI should start being the moral police. That’s what turns a lot of people off the FSF, after all.
“If you distribute any portion of the software in compiled or object code form, you may only do so under a license that complies with this license.” is quite odd wording, though I know what it means. It actually reads as though your binary license itself has to be distributed under the terms of the MS-PL, but that obviously isn’t case.
In my opinion, this isn’t the case. Your license for your derived compiled/binary work should only include the patent-litigation termination, no-trademarks and the AS-IS stuff. Both of them aren’t very hard to include in your final license. Now, the no-trademarks clause is meaningless in a binary distribution, while others are pretty common (well, maybe not the patent-litigation one.. but not so uncommon as well).
doesn’t the OSI have a rule against license profileration? I don’t get what the point of yet another license is.
when I first looked at MSPL, it seemed like a BSD-style licence with an added patent clause. examining it further, it looks like Microsoft’s own mix of BSD and GPL.
the similarities between MSPL and GPL is that code licenced under each can’t be relicenced. so if you make changes to MSPL licenced code, your changes will automatically be licenced under MSPL, same as with GPL. funny that Microsoft opted for the same ‘viral’ approach that they’ve continously attacked the GPL for using.
the differences are that while GPL requires the distributors to make the source code available
to end users, MSPL does not grant the end users these rights. so in that regard it can be employed in propriety solutions much like BSD code can.
sofar it’s pretty straightforward, however when looking at Microsoft’s overview regarding their licences they state that ‘Under the Ms-PL, licensees may change the source code and share it with others. Licensees may also charge a licensing fee for their modified work if they so wish.’
if I understand this correctly, it opens up the possibility MSPL code that has several contributors, each demanding a licence fee for use of their contributions.
anyway, IANAL, so my interpretations of this licence may be at fault.
funny that Microsoft opted for the same ‘viral’ approach that they’ve continously attacked the GPL for using.
This is *not* the same approach. Requirement to distribute under same license *IF* you distribute source code is needed to preserve original terms, expecially patents part, no-trademarks part and AS-IS part. For example, when you redistribute, can you let thirdy-parties use Microsoft’s trademarks? Of course not! This perfectly fits the spirit of a source-code distribution.
When you distribute in binary form, such ‘viral’ effect is not needed but the basic limitation (over original code) must be retained.
if I understand this correctly, it opens up the possibility MSPL code that has several contributors, each demanding a licence fee for use of their contributions.
Usually you have just one licensees even if code got changes from many parties. So I don’t think each contributor might ask you money (but they might ask money to licensee if it asks money from users, of course…).
Anyway, IANAL 😉
We need fewer licenses, not more. There’s really no need for other licenses than BSD, GPL, and one or two more.
We need fewer licenses, not more. There’s really no need for other licenses than BSD, GPL, and one or two more.
So, the GPLv3 is also not needed?
>>We need fewer licenses, not more. There’s really no need for other licenses than BSD, GPL, and one or two more.
>So, the GPLv3 is also not needed?
GPLv3 is needed but maybe now GPLv2 isn’t needed any more like GPLv2 made GPLv1 needless.
I agree with the previous post. If we look at it from a pragmatic point of view than there are 3 kind of Free Software licenses:
strong-copyleft, weak-copyleft and non-copyleft.
I think GPL, LGPL and BSD-License fills this spots quite well. Maybe we can than say that we need also for all three categories a license with more or less patent protection. Than we can add to the non-copyleft licenses the Apache license and maybe one or two more. So that we end up with: GPL, LGPL, BSD-license, Apache-license and maybe X and Y.
I know that there are a lot of egos out there who want to have their foo-license to immortalize them self into their license. But every pragmatic needs should be fulfilled with 3-6 licenses.
Edited 2007-08-19 12:18
How is GPLv2 not needed? 3 and 2 serve slightly different purposes now.
This is a case where choice is good. Even “too much” choice.
>How is GPLv2 not needed? 3 and 2 serve slightly different purposes now.
For me GPLv3 serves exactly the same purpose today like GPLv2 did in the past.
It gives all 4 freedom and makes sure that they will remain. But please don’t let this become another boring GPLv3 “good or bad?” discussion.
Edited 2007-08-19 12:57
It’s not a good or bad at all. 3 serves the same purpose as 2 for YOU, not for everyone. They are different. That’s a fact, not a discussion of good or bad.
>It’s not a good or bad at all. 3 serves the same purpose as 2 for YOU, not for everyone. They are different. That’s a fact, not a discussion of good or bad.
The purpose and the aims are entirely the same so if GPLv2 fits someone needs GPLv3 should fit them too.
Maybe someone has choosen GPLv2 to explore some holes which were created during the ~20 years of existence but that is something on which we can’t base our discussion.
But please stop this GPLv2 vs. GPLv3 discussion and focus on the important part of my statement above:
The the end there are only 3 kind of free software licenses: strong-copyleft, weak-copyleft and non-copyleft. So having 3-6 license should make everyone satisfied like i have already described above. There is no pragmatic place for more licenses.
“””
“””
Question: If you wanted that discussion to end, why did you contribute flamebait to it?
Some do not agree that GPLv3 is in the same spirit as GPLv2, and/or do not agree that it has the same aims. Note that tre question is not whether they do have the same spirit and aims, but whether significant code licensors feel that they do. And the answer is a very unambiguous “yes”. I can say that with confidence since many of them have come forward and been very direct on the matter.
That discussion is *very* relevant. There is a reasonable argument that the approval of GPLv3 would be contributing to open source license proliferation as much as the approval of any Microsoft license would.
Edited 2007-08-20 01:14
>Question: If you wanted that discussion to end, why did you contribute flamebait to it?
OK, here my previous post without anything about GPLv3. I hope now i will get some serious answers to the relevant topic of my message.
I agree with the previous post. If we look at it from a pragmatic point of view than there are 3 kind of Free Software licenses:
strong-copyleft, weak-copyleft and non-copyleft.
I think GPL, LGPL and BSD-License fills this spots quite well. Maybe we can than say that we need also for all three categories a license with more or less patent protection. Than we can add to the non-copyleft licenses the Apache license and maybe one or two more. So that we end up with: GPL, LGPL, BSD-license, Apache-license and maybe X and Y.
I know that there are a lot of egos out there who want to have their foo-license to immortalize them self into their license. But every pragmatic needs should be fulfilled with 3-6 licenses.
Edited 2007-08-20 01:16
I agree that, in general, GPL, LGPL, and BSD cover the spectrum pretty well.
I disagree that ego and immortality really play that much of a factor.
Fact is, everyone’s needs are different. Or at least, many *feel* that their needs are a little different.
It is worth mentioning that having a lot of licenses is not such a bad thing. It’s the copyleft licenses that make license proliferation such a negative thing. What people talking from the vantage point of GPL call “compatibility” is a misnomer. It means that code flow can occur in one way: other license -> GPL. As the whole issue of GPLv3 vs Apache licenses made clear, even that one way flow is tough to make happen, even with relatively permissive licenses. Even BSD was problematic for GPL at one time, until BSD changed to be *even* more permissive than their already right-next-to-public-domain status. Code flow back from the copyleft to the non is never even considered by copyleft advocates. (Of which I am one; I generally like copyleft.)
In my opinion, saying that “license proliferation is bad” is like calling *Windows* viruses *computer* viruses”. It obscures the fact that only one class of license experiences serious problems from it, and causes problems for other classes of licenses.
Edited 2007-08-20 01:38
>I agree that, in general, GPL, LGPL, and BSD cover the spectrum pretty well.
>I disagree that ego and immortality really play that much of a factor.
Than you have to explain why so much different licenses eixts which do basically the same just in their own words. E.g. why does we have a BSD License, MIT-License X11-License and some more which does at the end all the same thing with only some little different wording and under a different name? That’s just a example of a non-copyleft licenses. You can find also examples of weak-copyleft licences and strong-copyleft licenses. Most of this redundant licenses come also commercial free software projects/companies, which underline my presumption that they just want their “own license” and a licence with their name as an “addition” to their trademarks.
Rationally you can only decide if you want strong-copyleft, weak-copyleft or non-copyleft and pick one of the existence licenses. Their is no other argument than the “i want my own license” argument for all this licenses which differs only in some words to make it something new.
The purpose and aims are the same, but the details are slightly different. It is possible, and completely realistic, that someone does not want the clauses the GPL3 adds but everything else works for them, thus the GPL2 is the license for them and 3 is not. Likewise, if someone wants that extra “protection” the GPL3 provides, they go with it. v2 doesn’t provide them the same thing.
It is necessary for them both to exist.
It is necessary for people to be able to define their OWN terms for if they wish.
It is necessary for the OSI to exist to fairly and objectively define what is an Open Source license and what is not.
It is necessary for the freedom of choice to exist and for OSI to make it clear which licenses match the aims and goals of open source software.
Edited 2007-08-20 13:50
>It is necessary for the OSI to exist to fairly and objectively define what is an Open Source license and what is not.
For me this is not an important job of OSI. Independent from the OSI decision there are enough people and groups out their who can judge a license if it meets the Free Software Definition or the Debian Free Software Guide. Namely the FSF and Debian. They have already done this before OSI even exists.
An interesting position for OSI would be to use their “label” OSI-Approved to manage a small list of Free Software licenses that fullfill all pragmatic needs. Such a list could hold only 3-6 or maybe 10 licenses. This would be an important job because it would foster the license landscape and would help to lead people to good and often used licenses instead of exotic ones who often creates more problems than they solve (e.g. incomaptible licenses just because of some silly wording).
But OK, it all depends how you see OSIs position in the Free Software world.
But beside OSI, FSF and all other Free Software organisations i still believe that there is no pragmatic need for more than 3-6 licenses. Sure, you can always say “but i want foo and bar and therefore remove foobar” but this only creates endless (incompatible) licenses and, as i believe, often solves only the ego of some programmer/projects/companies.
It is rather easy to decide if you want to have strong copyleft, weak-copyleft or non-copyleft and than select one license out of this category. For all I care let it be GPLv3, GPLv2, LGPLv3, LGPLv2, Affero-GPL, BSD, Apache. But than i think you have covered all possible pragmatic options in Free Software licensing.
Edited 2007-08-20 14:14
So you care about “free software” and not “open source”. Fine, just say that orignally.
This article, however, is about open source.
On your other point.. trying to make a very small set of licenses fit everyones needs is.. very stupid.
>So you care about “free software” and not “open source”. Fine, just say that orignally.
>This article, however, is about open source.
Call it “Free Software” or call it “Open Source” that doesn’t make any difference for our discussion.
So can we go back to real arguments?
Edited 2007-08-20 15:31
http://www.gnu.org/philosophy/free-software-for-freedom.html
Not the same. It does matter. You seem to believe OSI is not fit for defining what is a “free software” license.
>http://www.gnu.org/philosophy/free-software-for-freedom.html
>Not the same. It does matter.
I emphasize different aspects of Free Software if i call it “Open Source” and arrange my values in a different order. But technically we speak about the same set of licenses. This shouldn’t be surprising because the “Open Source Definition” is not more than the renamed “Debian Free Software Guide” and the OSI calls them self a “marketing program for free software”.
>You seem to believe OSI is not fit for defining what is a “free software” license.
No, but i believe that with the label “OSI Approved” you could do better and more important things than just create a list of Free Software licenses.
But this is an entire different topic again compared to the topic about how many Free Software licenses are really necessary.
So before we haven’t finished one topic we shouldn’t jump to a different topic.
Edited 2007-08-20 17:05
you could do better and more important things than just create a list of Free Software licenses.
Why? Isn’t that their purpose? Isn’t that beneficial? One spot for anyone who is looking for a license to go.
I just don’t understand what points you are actually trying to make.
Devl wrote:
-“We need fewer licenses, not more. There’s really no need for other licenses than BSD, GPL, and one or two more.”
define need. if a developer doesn’t feel comfortable releasing his code under any existing licences why shouldn’t he be allowed to create a licence to support his needs?
at the end of the day, the licences that works for most developers will become the dominant ones. and those developers that choose not to use any of the ‘mainstream’ licences has every right not to do so. it may prove awkward to those who wishes to use their code but that’s life, beggars can’t be choosers. if you don’t want to comply with the conditions set by someone who’s work you want to use then that’s your problem.
diversity in licences reflect diversity in need and preference, I don’t see this as a bad thing.
What I was most surprised at when I first read this license text was, that there’s no fine print at the bottom stating that we reserve the right to change the terms of this license at any time without notice
or something similar 
So, there’s some hope after all
Microsoft will implement a opensource license on there terms just like a open file format that THEY want. I dont want a world where Microsoft dictates opensource and there version rather than already established good licenses.
If you can’t fight it, dupicate it with your own version, then let the version be the standard or dominate, it’s that simple.
“If you can’t fight it, dupicate it with your own version, then let the version be the standard or dominate, it’s that simple”
People can choose whatever license they want, there isn’t too much chance that MS’s license will dominate, seeing as most OSS projects will default to the GPL and other older, more politically correct licenses, this isn’t office suites, the people can and will choose what is best for them in this case.
Microsoft licenses will allow more developers to contribute the code. The developers who
– dislike GPL and similar licenses,
– don’t want to be any part of FSF dominated community
– are concerned about their IP
– are involved with Microsoft technologies
will feel safer and encouraged to share their code. As a result, FSF will have less influence among developers.
The license is clean and simple, easy to understand.
I believe that Bill and his team of millionaires have no concept of freedom of the market. Their model is based on total market domination not competition.
I must be missing something. If this license satisfies the OSI criteria, why wouldn’t it be approved? Not approving it for political reasons would completely discredit OSI.
Yes, Microsoft did use FreeBSD back in the day when it was better than Windows Server (in their minds).
I believe Microsoft used FreeBSD for their Hotmail and main HTTP servers, can’t be too sure, however.
Edited 2007-08-19 16:57
I argue that FreeBSD remained better than Windows Server all along
Back when Hotmail was HoTMaiL, it ran on FreeBSD; after MS acquired the service, they spent a lot of time and energy to migrate everything to Windows Server offerings of the time. Then they started touting it as a success case of migrating away from Unix to WS. And along came many of the exploits Hotmail became famous for
Very ironic, may I add
“I believe Microsoft used FreeBSD for their Hotmail and main HTTP servers, can’t be too sure, however. ”
Hotmail used FreeBSD when they were a separate company. When MS bought them, they eventually phased them out for Win2k and then Win2k3. MS never operated Hotmail on FreeBSD because they planned to, it’s because they bought it that way.
I found this in the article:
Code licensed under the MS-PL cannot be redistributed under the terms of any other license, but it can be combined with works released under the terms of other licenses, as long as those other licenses permit it.
How is that different from the restrictions GPL places on code? If people want to reject MS-PL using the above excuse, then the GPL should be tossed out as well.
Having pondered on this a while, read opinions from other folks, and tried to ignore the raging arguments from certain quarters about whether Microsoft is or isn’t immoral (seriously, set up a Jabber server or something to have that argument; nobody actually cares), there’s one aspect to the MS-PL that’s out of kilter with every other open source license that I’ve come across:
It’s a EULA.
It governs your use of the software, not what you do with it when you distribute it. Hypothetically, if you don’t agree to adhere to the conditions, you can’t even use it (even though the specific terms talk about distribution).
With the GPL—and most other licenses—you can be utterly and fundamentally opposed to the terms of the GPL, but still (perfectly legally) use the software because the GPL only covers distribution. If you want to distribute it (or a derivation) then it matters whether you adhere to the terms, and obviously you only can if you do.
In contrast, the MS-PL says:
This license governs use of the accompanying software. If you use the software, you accept this license. If you do not accept the license, do not use the software.
I’m not sure if there’s a specific reason for this, and it’s a bit of an odd one: obviously if you’re not planning on distributing the software and don’t agree to the terms then nobody would know if you were lying and went ahead and used the software, but other licenses don’t put you in that position in the first place.
Essentially, open source licenses don’t (usually) try to govern what you do with the software so long as it remains within the confines of your own hard disk.
Edited 2007-08-22 00:41