Conservancy at LibrePlanet 2018

by Molly deBlanc on March 22, 2018

March 24-25th hundreds of free software activists, community managers, hackers, legal experts, and all around fans will meet in Cambridge, MA for LibrePlanet 2018, the Free Software Foundation's annual conference and members meeting.

In addition to everything else happening there, Conservancy's own Bradley Kuhn, Brett Smith, Karen Sandler, and Denver Gingerich -- our part-time compliance engineer -- will all be speaking.

Topics covered include copyleft, the usibility of the GPL, medical devices, and free software in business.

• State of the copyleft union
  Bradley Kuhn, Distinguished Technologist
  

  The license-importance divide seems almost generational: the older generation cares about licenses, and the younger generation does not. Yet, the historical focus on licensing in FLOSS, while occasionally prone to pedantry to a degree only developers can love, stemmed from serious governance considerations regarding how community members interact.

  Copyleft was invented to solve the many problems of project governance, assuring the rights of users and creating equal footing for all contributors. The licensing infrastructure today also has increased in complexity, with proprietary relicensing business models, excessive use of CLAs, and tricky clauses on top of existing licenses.

  Given this climate, how do we understand if copyleft is succeeding? This talk explores historical motivations and modern reactions to these licensing matters, and digs into understanding how policies have impacted Free Software communities for both good and ill.

• A usability study of the GPL
  Brett Smith, Director of Strategic Initiatives
  

  We want software creators to use the GPL and its cousin licenses. We also know that people make mistakes in the process, or don’t even try because they’ve heard it’s "too complicated." Just as we do when we develop software, we would do well to study these failures and use them as opportunities to improve the usability of the GPL. This talk aims to start that process by identifying some known problems and considering some possible solutions. (None of these solutions are a new version of the license!)

• Copyleft, Diversity & Critical Infrastructure
  Karen Sandler, Executive Director
  

  GPL enforcement and Outreachy are the two most visible and controversial programs that Conservancy undertakes. In this talk, Karen will explore how the programs fit together in the context of software freedom generally. Karen will review her work around medical devices and critical infrastructure and show how seemingly disparate initiatives fit into a single advocacy narrative.

• Freedom, devices, and health
  Mad Price Ball, Rachel Kalmar, Dana Lewis, Karen Sandler
  

  When it comes to health, freedom is literally visceral. How do the principles of freedom apply to the devices used for medicine, health, and wellness? Moderated by Mad Price Ball, a Shuttleworth Foundation Fellow, this panel introduces leaders that bridge industry, community, and individual experiences. Rachel Kalmar (Berkman Klein Center), uses her experience with sensors and wearables to confront how devices and their data interact with a larger ecosystem. Dana Lewis (OpenAPS) connects us to health communities, and her work with the Nightscout project and patient-led efforts in type 1 diabetes. Karen Sandler (Software Freedom Conservancy) shares her experience as an individual with a device close to her heart: a defibrillator she uses, as a matter of life or death -- and she cannot get the source code to it. Join us to learn about how freedom matters for devices in health.

• In business: Keeping free software sustainable
  Denver Gingerich, Compliance Engineer
  

  Starting a business is a big decision, and choosing to share its results with the world is perhaps bigger still. Denver started JMP early last year, and faced this very choice, deciding to release all of JMP's code as free software and to charge money to use the instance he runs. In this session, Denver will describe why he chose to build a free software business, and will discuss the details of the business model he arrived at, alongside other business models for free software companies.

  Few contributors are paid to work on free software today, and far fewer are paid by non-profit organizations (or even by small businesses). It is imperative for us to explore how we can sell free software, especially through non-profits and small businesses, so we can bring freedom to more people and, just as importantly, build sustainable futures for our contributors.

More information is available on the LibrePlanet 2018 website.

Organized by the FSF, LibrePlanet is focused entirely on user freedom. We hope to see you there, in our talks or at the Conservancy booth in the exhibit hall.

[permalink]

Tags: conservancy, conferences

Copyleft compliance misconception #1: Companies check their source builds before publishing

by Denver Gingerich on February 15, 2018

We often hear from people that are confused about why companies fail to meet their copyleft compliance obligations - it seems fairly straight-forward to do, so why don't they all do it? In its many years of experience attempting to help companies comply with the GPL and other copyleft licenses, Conservancy has seen first-hand how many of the expectations software users have about how a company would release source tend to not be met the majority of the time. This post is Conservancy's first in a series on these common misconceptions about copyleft compliance, which will hopefully provide some insight for people wondering why these expectations are seemingly seldom met.

Misconception #1: Companies check their source builds before publishing

If you use or develop free and open source software, you probably find it natural for software projects to make building and installing their software as easy as posssible (or at least to provide contact points in case it is not). This is because getting people to use or contribute to such projects depends on these projects being straight-forward to build and install, otherwise people would just use something else (since normally they would have little invested in a project they can't build or install).

As a result, when companies publish source code as a result of their obligation to comply with the copyleft licenses in the software they distribute (usually this is their primary motivation - in rare cases (so far as we see it) companies are motivated primarily by engaging with the free software community, which we naturally try to encourage as much as possible), they do not have the same incentives as you would normally expect of a project distributing free and open source software. Consequently, companies tend to spend most of their time ensuring that whatever product they're selling to you (be it a router, Blu-ray player, smartphone, etc.) performs the functions it's intended to perform and meeting their bare obligations when it's shipped. They don't spend very much time working on the build and installation experience for those who would like to modify the software running on it after they receive it.

Furthermore, determining which parts of the device's overall build and installation process a company considers to be confidential is often not done, or is done under pressure so close to release time that the company does not wish to try untangling the portions they consider confidential from the portions that they are required to release in order to fulfill their obligations under the copyleft licenses for the software they chose to use.

What we normally see as the outcome of this (in the hundreds of source releases we've evaluated) is that the source code companies provide is nowhere close to meeting the requirements of GPLv2 (which is the copyleft license we most often see being violated), which state that companies must provide "the complete corresponding machine-readable source code" for all the GPLv2-licensed software on the devices they distribute and that "The source code for a work means the preferred form of the work for making modifications to it [which includes] the scripts used to control compilation and installation of the executable."

There are a variety of reasons that a company's source release might fail to meet the above requirements. In many cases we find that the versions of the source packages they provide are nowhere close to the versions of the binaries they distribute. Or their source release is missing entire source packages - i.e. they distribute a binary for Copylefted Project A, but do not provide any source code for Copylefted Project A, even though they might include source for Copylefted Project B.

Even if the above issues are not present, most often we find that there are no scripts at all for installing the software on the device (either in machine- or human-readable form), and any scripts we might find for controlling compilation of the executable are little more than what the original maintainers of the source package provided (which normally means people outside the company shipping the device). As a result, the compilation does not succeed because any changes the company made to the original source package are not considered in the build instructions (the company likely got it to build at least once, in order to ship it on the device, but decided not to document or share that process in their source release). This is particularly frustrating to the people who report the violations to us, as they often want access to the source code to do something particular with the devices they own.

After seeing this pattern in dozens of different source releases from dozens of different companies, it is clear to us that companies do not normally check to see if the source they release can actually be built and installed. Rather than being motivated primarily by meeting the perceived legal requirements of copyleft licenses and thus releasing markedly incomplete source code, our hope is that more companies will start to see the primary motivation for source releases as a way of engaging with the free software community, from which correct build and installation instructions (and indeed fully compliant source releases!) will naturally follow. We've already seen community development projects like SamyGo and OpenWRT form around Samsung TVs and Linksys routers and we hope other companies will see the benefits and help build such communities right from the start.

[permalink]

Tags: conservancy, GPL

Thank you to all our donors and Supporters - we did it!

by Karen Sandler on January 17, 2018

On behalf of Conservancy's staff and all of our member projects, I am excited to thank all of the people who contributed to this year's match challenge. Thanks to your generosity, we exceeded the amounts offered by Private Internet Access and an anonymous donor set for this year's annual fundraising drive.

What inspires me the most about this success is that we could not have done it without a high level of engagement from our volunteers. You not only donated your money to help sustain Conservancy, but you also took time to become a promotion machine for us. You blogged about it, you tweeted and tooted about it, you wrote about it on chat forums and you put up banners on websites. One volunteer even forwent payment on a small consulting gig and asked instead that the amount be donated to us.

Two years ago, we decided to become an individual supported charity to ensure independence from large corporate donors.Your support demonstrates that we can succeed and be vibrantly independent. We are humbled by your commitment to our mission and your trust in us and our work. We will use the money as best as we can to advance software freedom. We're so excited for the work we can all do together in 2018!

[permalink]

Tags: conservancy, supporter

Why Scènes À Faire Should Apply to Command-Line Interfaces

by Bradley M. Kuhn on January 3, 2018

Today, Conservancy joined other amici in the Cisco v. Arista case. Specifically, the amicus brief discusses why the scènes à faire affirmative defense for copyright infringement is appropriate and actually necessary regarding imitation of command-line interfaces. I hope this blog post will convince you that software freedom contributors should care about the issue.

The easiest example to understand these issues is Unix. Most of us know the basics of Unix's user interface, which primarily consists of commands that live in /bin and /usr/bin, that each include various command-line options that we've memorized. When the GNU project started, as RMS has described in his talks, he chose to imitate this user interface. Many reasons were obvious, but the most important one was that Unix was already an industry standard and users already knew its interface.

At the time, no one would have considered that you'd be liable for copyright infringement merely writing some new programs — 100% from scratch — that happened to have the same names and the same command-line options that were found in Unix. That interface, in fact, has been reimplemented at least a hundred times — by many Unix vendors and by various software freedom projects (GNU of course, but also by Conservancy's BusyBox project and others). As developers, we'd be incredulous if told that GNU infringed Unix's original copyrights. But that's exactly the argument that Cisco made about Arista's imitation of Cisco's command-line interface.

I'm not a fan of either Cisco nor Arista; all the software in question is proprietary software. Indeed, GitHub, which is one of our joined amici here, produces much proprietary software around Git, and that's bothers me too. I don't like it when any company writes proprietary software to work along with FLOSS. However, I agree with GitHub and Arista that copyright restrictions should not extend too far; copyright should not stifle simple command-line interoperatiblity. Merely imitating a command-line interface of one program in another should not cause (by itself) a copyright infringement.

Now, the last part to discuss are the questions: What is an affirmative defense, and what is scènes à faire? So, to explain it roughly with as little legalese (IANAL) as possible, an affirmative defense is one that you must prove after you're accused, usually through a trial (which is what occurred here). The burden is on the Defendant to prove that affirmative defense. (By contrast, if Arista had shown that, in fact, their command-line interface bore no similarity to Cisco's, that would have been a “negating defense”. Such defenses are much more assured to win, as they don't place such a burden on the defense.)

So, what, specifically, is the affirmative defense of scènes à faire? It's a concept originally from fictional works that generally expresses this idea: “if you're going to tell this story at all, you need at least these elements”. In this example, the analogy works like this: if your users will give a router textual commands via the command-line, that user will expect certain commands to work. Cisco's commands are industry standard and expected by users, similar to those in Unix. The amicus brief argues that this is a reasonable application of scènes à faire, because there is great benefit to the public and users if such imitation is permitted on command-lines without copyright restriction. Remember, under the USA Constitution, copyright exists as an “exclusive Right to … Writings” only because such exclusive controls “promote the Progress of Science and useful Arts”. Copyright is not an absolute right of control over written works by the authors, and its tentacles must be shortened by the public interest.

Finally, I call on the Linux Foundation to publicly ask their platinum member, Cisco, to stop this aggressive litigation on an edge case of copyright. Such a request would be consistent with the Linux Foundation's public criticism of others for copyright enforcement. This case is one where we all should stand together in the interests of free innovation.

[permalink]

Tags: conservancy, law