OSI Approves Microsoft Ms-PL and Ms-RL

Russ Nelson writes "In a board meeting held October 10th and announced today, the Open Source Initiative approved two of Microsoft's software licenses: the Microsoft Reciprocal License and the Microsoft Public License. These licenses are refreshingly short and clean, compared to, say, the GPLv3 and the Sun CDDL. They share a patent peace clause, a no-trademark-license clause, and they differ only in the essential clause of reciprocation. Of course, Microsoft is not widely trusted in the Open Source world, and their motives have been called into question during the approval discussions. How can they be attacking Open Source projects on one hand, and seeking not only to use open source methods, but even to use the OSI Approved Open Source trademark? Nobody knows for sure except Microsoft. But if you are confident that Open Source is the best way to develop software (as we at the Open Source Initiative are), then you can see why Microsoft would both attack Open Source and seek to use it. It is both their enemy and their salvation."

They're animals...

Captain James T. Kirk: They're animals.
Captain Spock: Jim, there is an historic opportunity here.
Captain James T. Kirk: Don't believe them. Don't trust them.
Captain Spock: They're dying.
Captain James T. Kirk: Let them die!

easy answer

"How can they be attacking Open Source projects on one hand, and seeking not only to use open source methods, but even to use the OSI Approved Open Source trademark?"

Extend, embrace.

Re:

For those who don't recognize the reference, here's a Wikipedia article on Embrace, Extend, and Extinguish [wikipedia.org].

Re:easy answer

The GNU project has tended to embrace and extend things. Not to the point of extinguishing them (except for classic UNIX, to a large degree.) But many people write shell scripts that traditionally run under /bin/sh but include bash extensions that make it impossible to run them under regular /bin/sh on other systems. The GCC contains extensions that don't deprecate and aren't flagged when the -pedantic switch is set. As a consequence C code is written that won't compile on other than GCC, and the developers aren't even made aware of this. The codebase slowly becomes GCC compilable only. All of this can be researched by anybody interested enough to look into it. The GNU project is an extend/embrace/extinguish operation, much like Microsoft.

Re:

Probably the exception that proves the rule, but I believe you can compile Bacula under MSVisualC, and MinGW (compile for Windows under Windows or Linux)

Also, all the Xen 'public' headers I have used so far under Windows compile under the WDK provided comp

gnu may embrace and extend but not extinguish

The GNU project has tended to embrace and extend things. Not to the point of extinguishing them.

Yeah, like almost any project that implements a standard. At some point they will add something... Ok...

The GNU project is an extend/embrace/extinguish opera

Yeah, one tiny little difference

GCC and Bash are opensource, you can get them to run on ANYTHING you bloody well want too. So if you want to use code that only compiles under gcc then all you need to do is get gcc to work on your system. Have you checked how many systems gcc works on? Go ahead, I will wait. Wow, long list eh?

Now compare this with closed source, lets say C#, how many systems does MS compiler run on? Oh, only windows. Wow that was quick.

Same with their IE, it ain't the problem that IE does things differently, it is that nobody really knows and can't copy that behaviour. THe problem ain't that IE does things differently as such, it is that they don't publish how to do it, so everybody else is left with browsers that run their "enhancements" slightly differently and end up with messed up pages.

All your post has done is to show WHY opensource is so essential. Frankly if this is the best attack you can muster against GNU, then we can sit back and relax, we won.

Re:easy answer

The difference is that GNU publishes their work, so while they extend other people's work it's usually to make improvements, and the improvements are always out in the open for other people to adopt.
If your compiler won't compile something but gcc will, you can easily work out what the gcc-specific extensions do either from documentation or source code, and implement them yourself.
By contrast, microsoft often keep as much detail about their extensions private as they can, to make it difficult for anyone else to implement.
Aside from that, gcc places far less other restrictions on you than microsoft products, you can run gcc on almost any platform, you can install it on as many systems as you want for no cost, you can modify it to suit your needs, you can redistribute it with very few restrictions.

GCC then and now

Back when GCC was managed by RMS, he liked adding "convenient" extensions to the language. One of the reasons was to "tie" the programs to GCC as you describe.

The current maintainers have quite a different view on them, they only add extensions for things that can't be expressed in the language (mostly stuff close to hardware, or sometimes optimization hints), and they give them __unwieldy_names__. And they are slowly removing the convenience features.

If you don't get a warning from an extension (rather than a "quality of implementation" issue) with -ansi -pedantic it is a bug.

There probably are such bugs (GCC is big and complex, and there are stuff in the language that can't really be tested) but I guess you can't actually mention any, since you are most likely just a troll making stuff up as you go. So I challenge you to mention one such issue that can not be found in the bug database to prove you are not a troll. And if you can mention two such issues, I'm willing to believe that you are not a kook either. In either case, please submit a bug report when done :-)

Re:easy answer

It's decided by a very small number of the community. The fact that the rest have reasons not to help (whether it be apathy or a time/resource conflict) doesn't negate the impact upon them.

Microsoft represents a small number of the community as well. The analogy is pretty sound. I don't think the point to be taken is that GNU is evil, though. It's that extension is how software evolution works, and a significant number of people will migrate to the most refined and feature-rich solution, sometimes causing the previous generation of product to die.

The problem with Microsoft isn't really the embrace/extend part. It's the extinguish part. While bash pretty much replaces sh, you can pretty much count on bash to always be around until something replaces it. The FSF isn't going to just kill it one day so it can cannibalize the market.

Re:easy answer

The problem with Microsoft isn't really the embrace/extend part.

Kerberos
Java
HTML
RSS
MAPI
Etc, etc...

Re:easy answer

Same way Sony can tell people that even copying a CD for use on a personal MP3 player is pirating, while at the same time selling VAIO computers that rip CDs, and portable music players that come with software that helps you rip CDs, as well as selling CD burners, and blank disks to copy CDs onto. They are such a big company, that the left hand doesn't know what the right hand is doing. What's good for one department isn't always good for another department, and sometimes the needs of different departments within the same organization clash.

Re:easy answer

Microsoft wants all FOSS to be developed for Windows. And Office. As long as it does not compete with their products. Next comes the extend piece, in which they fiddle with someone's creative effort just enough to come up with something they can protect, like how they did with kerberos and ldap. Free development for their profit.

It's about patents

This is a way for MS to get some amunition against patent litigition. Just release the next Windows SDK with some core code licensed under the new MS-PL, make a huge "MS now Open Source" new splash, and wait until the first company that has a Windows produ

Re:

Well, I am pretty sure this is not a CRC error this time. Sounds more like they are applying quantum computing to their business model. Now, they can make all claims simultaneously, claim they are anything and everything and never have to settle on state

Not OSL.

I can't see how a license that governs use rather than distribution can be considered open source.

I'm dissapointed in the OSI.

Re:

I can't see how a license that governs use rather than distribution can be considered open source.

You mean like GPLv3? "You can't use this code to build a TiVo." (My paraphrase.)

Re:Not OSL.

You can use GPLd (any version) software to build anything you like. The license doesn't come into play until you distribute the software.

Why do people find this so hard to understand? Its a simple concept.

Re:

"You can use GPLd (any version) software to build anything you like. The license doesn't come into play until you distribute the software.
Why do people find this so hard to understand? "

Because GPL is overly complicated.
As said by Russ Nelson (quoting him

Re:Not OSL.

Bull. Google runs a customized version of Linux and only uses it internally, they aren't required to give the source to people who visit their website. Plenty of people/companies go to the expense of building internal-only software.

Re:

Maybe this commercial company wants to have their homebuilt TiVo in the breakroom. Maybe it's not a TiVo, it's a cash register, and they want to use it, not sell it. Maybe commercial companies have software that they use, rather than sell.

If there were w

You mean like the apache license?

From the Apache License [opensource.org]

TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document.

Not only is this license approved by the OSI as open source, it is also considered to be a Free Software License by the FSF and is even compatible with GPLv3.

Re:You mean like the apache license?

Yes, but the apache license puts no conditions on use, just grants extra rights(ie to patents held by Apache developers).

The MS licenses puts conditions on use - for instance, granting MS permission to use your patents.

Re:You mean like the apache license?

The MS licenses puts conditions on use - for instance, granting MS permission to use your patents.

I'm not seeing that. Section 2B states that each contributor to grants access to any patents relevent to his contribution. However, that only applies to recipients of the software, and thus is irrelevant until you distribute your modified version. Furthermore, if you are simply a user, not a contributor, then you are not granting patent license to anyone. Lastly, even when you contribute and distribute, the patent license is only relevent to your contribution and derivative works thereof - not to unrelated projects that Microsoft may develop.

Now that you got me thinking though, Section 3B may be a restriction on use, and can apply even if you don't distribute the work. However, one could argue that if you are suing someone about a patent, it is because you don't consider it to be valid. Thus you aren't forced to stop using the software, and if you win the court case you will have done nothing illegal by continuing to use it. Furthermore, recent supreme court rulings state that continuing to allegedly infringe upon a patent while you are challenging it is not grounds for willfull infringement. Regardless section 3B is practically identical to the patent litigation clause in section 3 of the Apache License (v2).

Granted, the wording of the introduction to the MS license(s) is much more forcefull than the Apache License, and its similarity to an EULA did give me pause at first. But after reading it, I don't think it restricts use anymore that the Apache License (if at all). If you are simply using the software, then the license merely requires you to agree to not distribute contrary to the license, which you aren't doing anyway. If you are going to redistribute it, what difference does it make whether you agree to follow the license when you start using it or when you start distributing it?

Re:

So it should be moderated "Wrong" not "Troll".

Oh wait, there isn't a "Wrong" moderation, you're supposed to reply if something is wrong not just mod it down.

Re:

I think one of the most stringent set of license requirements in the community is the Debian Free Software Guidelines [debian.org]. After reading the MS-PL and MS-RL, it seems that it would meet the DFSG. I'm sure there will immenently be debate on the debian-legal l

Interesting Years Ahead

This is going to be quite interesting to see. One can hope that good things will come of this, but I honestly don't see it. Luckily, I'm not in a position whereby my mistrust will affect the outcome negatively.

How? Simple

How can they be attacking Open Source projects on one hand, and seeking not only to use open source methods, but even to use the OSI Approved Open Source trademark? Nobody knows for sure except Microsoft.

They have more than one bit in their brains to make decisions. Hence "open source" is not a knee jerk reaction to them, in a way that "Microsoft" is a knee jerk reaction to certain people in the community.

Open Source is a model, it's a tool, to achieve a purpose. A serious company doesn't shy to use the tools at its disposal, even if some simpler folk might find this contradictory upon first sight.

Re:How? Simple

It's not a 'someone'. Microsoft has 80k employees. No doubt some of them would like to use any tool (even a chair) to beat you senseless and some who no doubt would share their source with you without a second though. The fact that Microsoft has enabled the latter is, at least somewhat, commendable.

Dang literalists. /me sighs. Fine. Take 2, for Skippy here.
Old and broke:

I have no problem with someone learning to use a hammer. I DO have a problem with it if the only reason they're learning to use a hammer is to beat me senseless with it.

New hotness:

It's {arguably} laudable that Microsoft would create these licenses. However, given their present demeanor towards FOSS, the timing of such work in this area could be construed as somewhat suspect.

Howzat?

How could this get approved?

The "license" states in bold letters:

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.

This is contrary to any Open Source license I know of. The whole point of Open Source is that you can use the software in any way you want. You have to agree to the license only when you distribute. Microsoft is attempting to subvert OSI, just like it has already subverted ISO.

Re:How could this get approved?

This is contrary to any Open Source license I know of.

There are at least 2 that explicitly say that .. Common Public License [opensource.org], Eclipse Public License [opensource.org]

Re:How could this get approved?

This is contrary to any Open Source license I know of. The whole point of Open Source is that you can use the software in any way you want. You have to agree to the license only when you distribute. Microsoft is attempting to subvert OSI, just like it has already subverted ISO.

You may be confusing Open Source with Free Software [gnu.org]. Free Software guarantees the freedom for the user to use the software in any way they feel--Open Source does not. At least Open Source in the sense that I know it, does not guarantee that particular freedom [gnu.org] and is the reason Microsoft can push supposed open source licensees that restrict user freedom.

Re:

and is the reason Microsoft can push supposed open source licensees that restrict user freedom.

As opposed to the FSF pushing licenses that restrict developer freedom?

I am not sure what gave you that impression.

Interestingly the OSI's Open Source Definition is derived from an attempt to define the FSF's 4 freedoms.

The lineage goes from the FSF to Debian (and the Debian Free Software Guidelines) to the OSI's Open Source Definition (which were mostly copied from the DFSG).

If you read the MS-Reciprical License and the MS-PL, you will see that they don't provide any restrictions on use, so this distinction doesn't really matter anyway.

I would also note that the FSF

isn not exactly the paragon of Freedom as in Free Speech either. As much as Debian may be a dysfunctional democracy, they are better at consistantly addressing this issue than the FSF.

Ever wonder why the GFDL has the invariant sections clause? According to Stallman himself (in a post to debian-legal), it was because he wanted to *force* the distribution of the GNU Manifesto with the Emacs manual. This essentially turns the ideal of free speech on its head by creating a situation where forced advocacy is accepted. When the then-main-architect of HURD criticized the decision, Stallman asked him to resign. If this is the sort of Stallmanist "Free Speech" we are to associate with Free Software, I want nothing to do with it.

Debian did the right thing and to this day considers any GFDL work containing invariant sections to be non-Free.

Note that I am not a Debina Developer, and I think that at the end we should think for ourselves and not be groupies of RMS or anyone else.

Damn, they actually look reasonable

The Ms-PL looks basically like the same terms as the BSD/MIT license with a patent peace clause, and the Ms-RL looks like the same thing with a Mozilla PL-like reciprocal clause. Neither one looks like the GPL. That's an unalloyed good thing.

Re:Damn, they actually look reasonable

1) These licenses aren't substantially the same.

2) There's a risk in writing a succinct license in that you might not cover each and every case in detail. Yes, by being so succinct, Microsoft is taking a risk that the judge might not accept the facial meaning of each clause.

3) Among other things, they're short and readable. Specifically, they don't name a jurisdiction. This is VERY GOOD for international projects. How would you like to have to sue somebody in Santa Clara simply because you contributed to an MPL-licensed code and they infringed it?

4) Why do you think you can't trust us? If you think the licenses don't comply with the Open Source Definition, you should say exactly why, rather than attempt to raise FUD. As [e-mail] I can definitively state that your email address above is not on the license-discuss mailing list -- if you don't participate in the process, why should anybody believe your criticism of the result?

Re:

You normally are not allowed to distribute other peoples code without a license, so you NEED to agree to a license to do so. (GPL)

You ARE normally allowed to use software you have bought, so you can do NOT have to agree to a license to do so. (EULA)

This li

Market control, but the possibility for change

Microsoft is a marketing company. They would much rather have Microsft Reciprocal Licensed Linux VistaXP Edition than GNU-Linux. They can always sell their branded products to clueless managers while scaring them with patent threats against the other guys.

Microsoft is also a follower. They typically are behind the bleeding edge of technology but always attempt to buy up as much as possible and claim they "innovated" it. Microsoft research is a big exception, and I wouldn't doubt that it's people from Microsoft Research driving the new licenses as well.

Microsoft is also not entirely stupid; they intend to attract developers with their licenses not to release any Microsoft products under them, but to bring open source development onto Microsoft platforms. That, ultimately, is a war that open source can only win by having a fundamentally better product. If Microsoft opens its documentation and internals to developers, most of them will see fewer advantages in pure open source/free software systems. All Microsoft has to do to keep making money is keep Windows shipping on every PC shipped in the world. Even if they're forced by the market to open source all of Windows, they will still own the trademarks (and patents) and probably still ship Windows on a significant number of PCs. Most home users don't give a shit what they run and will happily buy Microsoft, especially if their formerly Linux-enthusiast friends now see Microsoft as an open source company.

In the end, cooperation really is the goal of free software/open source anyway. Destroying Microsoft would be a net loss for everyone. Microsoft slowly converting to an open/free development model is a scenario in which everyone wins. Who will care if they run Linux or Windows if both support Posix and Win32 environments using the best elements from each kernel? To be honest I don't think this is very likely; it's much more likely that the Next Big Thing will push the operating system question into the realm of questions like which general purpose sorting algorithm is the best.

"They" is rarely a single viewpoint.

How can they be attacking Open Source projects on one hand, and seeking not only to use open source methods, but even to use the OSI Approved Open Source trademark?

How can Americans be refusing to sign up to Kyoto on one hand and winning Nobel prizes on the other? Probably because Bush isn't Gore and Gore isn't Bush. When it comes to human rights abuses, the administration can be winking when telling people not to do anything bad and yet you can still have a general who comes down like a ton of bricks on any soldier who even begins to act out of line because he feels another path makes more sense in the long run.

If many of the old guard senior execs feel one way - and a newer junior VP who has his senior VP's protection feels another - then it's entirely possible for two parts of a large organization to act in two apparently conflicting ways.

That's simply the nature of large organizations. Once you clear a certain size, you can't have every decision cross your CEO's desk or they'll get nothing done.

Attribution clause

It's interesting the attribution clause included on these licenses. We have many discussions about how manage attribution on open source projects. CPAL (https://www.socialtext.net/open/index.cgi?cpal) license gave us some points to consider when talking about attribution. This license allows you to put a limited attribution on the software "copyright notice, short phrase (10 words), graphic image and URL." GPL 3 has it's own clause on attribution resources (see clause 7) The way Ms-RL Ms-PL describes how attribution will be managed has another aproach from the previous licenses: They require you to respect any attribution notices that the original work has. So here we have an OSI approved license which can be used to preserve all your attribution notices...

So, now that they have these licenses...

does anyone have any guesses as to what software/code they will be willing to release under these terms?

Re:

IronPython, IronRuby, .NET's Dynamic Runtime Library (DLR) (used by .NET and Silverlight) are among the code that Microsoft has already released under Ms-PL.

definitions

Ms-PL: BSD-like
Ms-RL: GPL(2?)-like

It'll be interesting to see what the FSF has to say on these.

Because Microsoft is not a person

How can they be attacking Open Source projects on one hand, and seeking not only to use open source methods, but even to use the OSI Approved Open Source trademark?

This is because Microsoft is not a person, it is a corporation which is a legal fiction enabling a company to be treated as a single legal entity. In reality, Microsoft is ~100,000 people who have their own ideas, ambitions and goals. As an employee, I think my ideas of what I want to do vary significantly based on the number of levels I have to go up to find a common manager with someone else in the company.

Not even MS knows why...

How can they be attacking Open Source projects on one hand, and seeking not only to use open source methods, but even to use the OSI Approved Open Source trademark? Nobody knows for sure except Microsoft.

No, not even Microsoft itself... MSFT is a huge corporation. Nobody, not even Ballmer and Gates controls or dictates fully the company actions. Many people inside the company might have different motives to push for OSI. Maybe some push it because they (in some way) believe in open source. Maybe some push it because they just want to give MSFT a better image. Maybe some think its a way to fight open source. Probably some other though it is a bad idea...

clean... or simplistic?

The FSF and other organization writing open source licenses don't deliberately complicate licenses. The reason their licenses are long is because the legal environment is complex and there are a lot of cases they need to cover.

You may prefer a "refreshingly short and clean" library, but only until you have a disturbingly lengthy and messy debate with a lawyer or a court about what the intent and meaning of that license is and the court decides that people could reasonably interpret the license differently from the way you intended. The appearance of simplicity and actual simplicity are two different things, in licenses as much as in software.

I think it's unlikely that Microsoft's licenses contain a deliberate legal backdoor. And OSI approval is a good first step. But those licenses are unproven in practice, have no history behind them, and haven't been analyzed carefully by a lot of people. Maybe they'll eventually turn out to be reasonable and sound, but for now, I'd stick with one of the proven open source licenses.

Ms-PL: The trap here seems obvious...

The license reads "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."
This limits non-copyright rights by tying usage to acceptance of the copyright and patent right parts of the license. I wouldn't use it for that reason alone.
Here's an example: Ms-PL Product A is a flower shop management tool. OL (Other License) Product B is my solely-developed tool that controls plastic forming robots in my flowerpot factory. There's no code commonality between them, including no shared patent technology.
I don't add ANYTHING AT ALL to product A, but I use it and sell it to my customers. MS finds a way to apply product B's patented tech to product A. So MS steals product B's patented tech, and adds it to product A. I sue MS for infringing Product B's patents. MS automatically has the right to countersue me for ALL of MS's patents in product A and any A-derived product, including legal MS patents I'm otherwise in compliance with. I can't just stick to the pre-infringement version of product A; I must use a version of product A that has NO MS technology in it either, or not use product A at all.
This allows MS to shut down competitors whose only "violation" against MS's legal patents is a dispute over the competitor's patents in other products. This multiplies the patent snarl that's already threatening the industry.
Furthermore,"2. Grant of Rights (B) Patent Grant-... each contributor grants you a...license..to ... use, sell, offer for sale, ...contribution in the software or derivative works of the contribution in the software.
.... 3. Conditions and Limitations (B) If you bring a patent claim against any contributor over patents that you claim are infringed by the software, your patent license from such contributor to the software ends automatically."
It also seems to mean that anywhere a patent goes, this license applies.
If MS takes its legally patented tech from product A and adds it to product C, then the patent snarl described above will also deprive you of product C. A few carefully crafted patents, spread around like paint, and MS destroys all competitors once and for all.

Re:

Except open source isn't one of those things that can be extinguished.

I agree with you, but all it takes is one sentence with tricky wording to cause a bunch of trouble. My guess would be that this is where they're going with it.

A one page, understandable OS Licence?

I think the Microsoft licences are really impressive. Short, easy to read, and it seems obvious to me why they were approved. These Microsoft licences would be a good first read for someone trying to understand more sophisticated licences, like GPL v2 or v