If the market's response to the Affero GPL is any indication, I was 100 percent wrong to suggest that open source would suffer without closing the so-called "ASP loophole."
That, at least, is the feeling I'm getting reading Stephen O'Grady's excellent summary of open-source licensing, and particularly the GPL, and how it works (or doesn't) in SaaS, cloud, and other instantiations of network-based computing. Despite the fact that the Open Source Initiative approved the Affero GPL--which explicitly shuts the door on free-riding on open source in network-based computing without contributing back--few have adopted it.
This could be because we need to raise awareness of the AGPL. Or perhaps it means no one really cares.
Yes, Fabrizio Capobianco, a personal friend and CEO of Funambol, an open-source mobile company, is right to suggest that Google has profited handsomely from open source while giving commensurately little back, but I'm starting to wonder, along with O'Grady, if it matters. General Electric uses Alfresco's software throughout the company while paying us nothing...and yet we're having a banner year.
Perhaps this is just the cost of doing business in open source? I definitely believe that open-source companies derive far more value from free distribution than we lose.
I also believe that Google contributes open-source code where and how necessary for it to compete effectively. Google has become very active in a wide range of open-source projects. Perhaps it's less important to worry about its paltry contributions back to Linux, and instead take a more holistic view?
And perhaps the lack of uptake on the AGPL is a recognition that the value in open source is less about contribution and more about distribution and adoption. The reality is that very few open-source projects can command any outside contributions of note, but many are able to become widely used by lowering the barriers to adoption through free downloads and light restrictions on use.
In sum, perhaps Richard Stallman was wrong. Perhaps open source's cardinal virtue is not freedom to modify source code, but to get it in the first place.
Follow me on Twitter @mjasay.
I read Richard Stallman's commentary on cloud computing in the UK's Guardian. Stallman is full of warnings about cloud computing:
One reason you should not use Web applications to do your computing is that you lose control. It's just as bad as using a proprietary program.
But he completely neglects to mention that he had a chance to seed the cloud, which is largely built using open-source software, with an upgraded GNU General Public License (Version 3), and he demurred. Richard Stallman and the Free Software Foundation failed to protect the cloud when they had the chance, finally capitulating to industry pressure with the Affero GPL, an inelegant but workable sop.
Ars Technica is right to suggest that Stallman's doom-and-gloom about cloud computing is "myopic" and utters a final judgment well before even a provisional judgment is warranted.
But Stallman's biggest fault here is that he criticizes a problem that his own inaction created.
First it was the Affero General Public License that Google banned from its Google Code site, an open-source code hosting site. Google contended that it didn't want to encourage license proliferation by accepting projects using licenses that don't have widespread use and acceptance.
This week, however, Google nixed a highly popular, important license license: Mozilla Public License.
Google's Chris DiBona played the proliferation card again against the MPL, but also admitted that how Google determines whether a license is suitably popular is "so arbitrary." Great. That makes me feel better. At least there's a clear criterion for deciding. Not.
While some projects have moved away from the MPL in recent years, it remains one open source's standard licenses. I've got to think this has more to do with MPL derivatives (It's no secret that DiBona disliked the "badgeware" licenses that derived from the MPL) and their potential impact on Google's ability to consume their code, just as with the Affero GPL, than with any respect for license proliferation.
If it were about proliferation, Google would settle on GPL/LGPL, BSD/Apache, and MPL. Between those, most licensing preferences would be covered. By leaving out the MPL, however, Google has mistakenly dumped the baby with the bath water.
I and others have argued that it's critical to open source's future that licenses like the Affero GPL close the "ASP loophole" by requiring companies like Google to contribute back derivative works of open-source software that they distribute as a service, rather than as packaged software. Now Gordon Haff is suggesting that requiring Web 2.0 to Contribute 1.0 may cause more problems than it solves, and he could well be right.
The problem has nothing to do with whether Web 2.0 vendors like Google are required to contribute back. The problem is all the so-called Web 2.0 users:
Distribution in the GPLv2 and GPLv3 licenses draws (mostly) a hard-edged line. If you're an enterprise using software internally, anything goes. If you're using GPL code in software you're selling to the public--whether downloaded, on a CD, or in embedded firmware--you must make the relevant sources available. However, as more and more companies of every stripe make parts of their computing infrastructure available to their customers--think online banking, for example--where does it end? The boundaries become very fuzzy--which would inject lots of uncertainty into just about any use of open source in an enterprise environment.
This is a very, very good point. I'm not sure how to answer it.
... Read moreI've written about Launchpad, Ubuntu's software hosting and development website that enables collaboration across multiple projects, but I'm even more excited now that Mark Shuttleworth is strongly considering releasing it under the AGPL (Affero GPL). Launchpad is very cool. Keeping it open in a networked world makes it even cooler.
The choice of AGPL - which specifically covers software offered as a networked service - would be appropriate for Launchpad. It would also add some much-needed credibility to AGPL, which has come in for criticism from Chris DiBona, Google's open source program manager. DiBona has said he wants to see more examples of AGPL in action before deciding whether Google supports the license. Google has closed the door to AGPL on its Google Code host site, forcing projects that use the license to leave.
Indeed. Ubuntu is about as close to the cool kid at school as one gets in open source, and Launchpad is at the heart of Ubuntu. This would be a big win for the AGPL.
This post from Pierre Teilhard de Chardin is so funny that it's almost an instant open-source classic. Pierre sets up Google's (mis)statements on the AGPL and its licensing beliefs against one another, leaving a rather ramshackle mess.
It's all in good fun, and in Google's defense Chris Dibona has stated that if AGPL were to become as popular as GPL/BSD/etc. then Google will start hosting those projects. He doesn't like AGPL but isn't blocking it "for nefarious reasons." Fair enough. In the meantime, enjoy Pierre's post.
(By the way, one question I have is how to measure popularity. There's volume of projects, of course, but there's also importance of projects. If Linux were to go AGPL, and it were the only one, wouldn't that be enough? And since when is Google the arbiter of what counts as open source? I thought that was the OSI's job?)
Google apparently likes open source that lets it "borrow" open-source software while giving comparatively little back, and always on Google's terms. While I think Google has been doing better of late vis-a-vis open source, its policy of blocking projects from its Google Code forge that are licensed under the AGPL is wrong and a betrayal of the open-source principles it claims to respect and approve.
As Google's Chris DiBona says,
In fact we do not support the AGPL on code.google.com....It is also not okay to host an AGPL covered program on code.google.com by saying it is GPL, as you are telling the users of the site one thing, while meaning something else altogether. So sadly, the answer is to remove your project and host somewhere else like sf or savannah.
Well, no, Chris, AGPL is not "meaning something else altogether." It actually means precisely what the GPL was always intended to mean: Reciprocity. It is likely true that Google doesn't like that reciprocity requirement, but that's "something else altogether."
What is the AGPL? It's the Affero General Public License, and finishes the job that GPLv3 was supposed to do: Broaden the definition of "distribution" enough to keep Web freeriders like Google, Digg, etc. from using open-source code without contributing back.
... Read moreIn my estimation, the GPL is by far the world's best open-source license for business. If Palamida's recent count of GPLv3-licensed projects is any indication, business in open source is very, very good:
Our database now contains over 2,000 projects that are using the GPL v3. At this rate the GPL v3 is being adopted by 1,000 projects every 4-5 months, and if the trend continues, the license will be used by 5,000 projects by the end of the year.
It will take time for GPLv3 to achieve the same level of trust that GPLv2 has enjoyed, but at this rate it may be happening sooner rather than later.
In the meantime, Palamida, a request: Could we get a count on the number of projects using the Affero GPL (AGPL)? It will be miniscule today but I'd be interested in following its growth.
One of open source's biggest failings has been to extend its relevance into the Software as a Service world. The OSI has finally corrected this with the approval of the Affero GPL.
Fabrizio Capobianco, CEO of mobile open-source company Funambol, has been the most ardent crusader for development and approval of a license like the AGPL. In a blog posting, he talks through the importance of the AGPL, and identifies perhaps its biggest opponent: Google.
In GPL v2, those who ran open source software in a Software-as-a-Service (SaaS) environment, and modified the open source code, were not required to return the changes back to the community....For me, this has always been one of the worst risks for open source oblivion. If you can take and you do not give back, defeating the copyleft concept, you kill open source. The ASP loophole is the cancer of open source.... ... Read more
- prev
- 1
- next
