Choosing a License for Mailpile 1.0
Posted by Bjarni on May 8, 2015
Dear Mailpile backers and supporters!
The time has come for us to make a decision. Back when the project first got off the ground, we had some difficulty deciding which Open Source license we should use for our work. We narrowed it down to two licenses:
The Mailpile development code, the Alphas and the Betas have all been released under both of these licenses. This was a legal hack to avoid having to choose right away, without having to ask contributors to sign copyright waivers before sharing code with the project.
However, before the 1.0, we need to make a decision and settle upon one license or the other.
We have decided to let our community decide, and are in the process of sending out invitations to our community governance page.
Why do licenses matter?
Although a bit boring, licensing is actually very important.
The license is what grants exemptions from the rule of international copyright law. Without a free and open license, our backers would not be guaranteed the right to download, use, modify and share the results of our work - the work they paid for.
The license also helps define the community around the software, and this is where things get tricky.
Some licenses, especially those championed by our friends at the Free Software Foundation, are designed to protect the rights of the end-user and ensure that community efforts remain with the community. The AGPLv3 is the most modern and powerful such license, designed for software like Mailpile.
Other licenses (such as the Apache license) are more permissive, allowing corporations to make use of community code in traditional, commercial software products. This fosters collaboration and motivates companies to contribute valuable time and effort, but does so at the expense of end-user freedom since many users of the software will end up using proprietary derivitives.
Permissive licensing: Webkit
An excellent example of permissive licensing, is the Webkit project.
Webkit is based on KHTML, which was published under a relatively permissive license (the GNU Lesser General Public License). I say relatively permissive, because although the LGPL allows use within proprietary products it does require that motifications to KHTML itself be given back to the community. This balance allowed Apple to create Safari, a proprietary browser, while forcing them to publish the engine they derived from KHTML - an engine which we know today as Webkit.
Webkit was then chosen by Google for use in their Chrome project, and the rest is history: Webkit, daughter of KHTML and Apple, has taken the lead as the world's fastest and most powerful HTML5 rendering engine. It is now not only at the heart of both Safari and Chrome, but is used in a wide range of other smaller browsers and open source projects.
This is a fantastic open source success story, and it's worth remembering that this would probably never have happened if KHTML had been licensed using a more permissive license, such as MIT, BSD or Apache - those license would have allowed Apple to keep their improvements to themselves.
But there are more lessons to be learned here!
Today Safari and Chrome are both provided free of charge, but they are proprietary programs and the community does not have access to their inner workings. Amongst other things, this means the community cannot easily inspect or audit how and when these browsers "phone home" and share private data with the companies that wrote them.
It can also be argued Webkit's license perpetuates the status quo and unfairly skews the market against truly free software. Compared with these truly free alternatives, Chrome and Safari will always have a head start because Google and Apple can keep some of their innovations to themselves.
Of course the counter-argument is, that without the motivation for big companies to collaborate, Webkit probably wouldn't exist at all.
What about Mailpile?
Although there are lessons to be learned, Mailpile is different from Webkit. Webkit is a technical tool, one building block of many that developers use when writing a browser. Mailpile on the other hand, is a stand-alone end-user application, and (when it is finished) Mailpile can be used directly to provide service online.
Theoretically, a company could use Mailpile's code to build a service which competes directly with GMail and Hotmail - but if Mailpile's license does not require they share their changes, the users of these competing services will be no more free than they were before. There is also no guarantee that these companies will contribute to the community version of Mailpile.
On the other hand, many companies (and some individuals) strongly dislike the AGPL and avoid all software which uses it. These people will be excluded from our community and will not help us improve Mailpile if we use that license.
So that becomes the crux of the question.
Do we want to compel companies to share their changes to our code, which is what the AGPL does, or do we want to take a more trusting approach and rely on enlightened self interest and good-will - by releasing under the Apache license?
For some, this is an idealogical question, a matter of right or wrong. Is it morally acceptable to allow people to benefit from Mailpile's work without contributing back? Is it morally acceptable for Mailpile's authors to tie the hands of other businesses?
For others, this is merely a matter of strategy and tactics. What is most likely to improve end-user freedom in the long run? What is most likely to help Mailpile itself develop and thrive? Will chosing a restrictive license marginalize the project and prevent us from affecting wide-spread improvements to how people use e-mail? Will a permissive license make it easier for abusive companies to launch new webmail services as honey-pots to harvest and collect private e-mail data?
Those are the questions we need to answer!
If you have an opinion, one way or another, I encourage you to write blog post explaining your position (or send links you feel are relevant). If you prefer to publish yourself I'll be happy to link to your website - or we can publish guest posts here on this blog.
Hopefully a few people will express their opinions and help our community vote one way or another.