Congratulations to Tesla on Their First Public Step Toward GPL Compliance

by Bradley M. Kuhn and Karen M. Sandler on May 18, 2018

Conservancy rarely talks publicly about specifics in its ongoing GNU General Public License (GPL) enforcement and compliance activity, in accordance with our Principles of Community Oriented GPL Enforcement. We usually keep our compliance matters confidential — not for our own sake — but for the sake of violators who request discretion to fix their mistakes without fear of public reprisal. As occurred a few years ago with Samsung, we're thrilled when a GPL violator decides to talk about their violation and works to correct it publicly. This gives us the opportunity to shine light on the real-world work of GPL and copyleft compliance.

We're thus glad that, this week, Tesla has acted publicly regarding its current GPL violations and has announced that they've taken their first steps toward compliance. While Tesla acknowledges that they still have more work to do, their recent actions show progress toward compliance and a commitment to getting all the way there.

Conservancy has been engaging with Tesla on its GPL compliance since June 2013, when we advised Tesla that we had received multiple reports of a GPL violation regarding Tesla's Model S. Customers who purchased Tesla's Model S received on-board system(s) that contained BusyBox and Linux, but did not receive any source code, nor an offer for the source. In parallel, we also asked other entities to advise Tesla about GPL compliance. We know that Tesla received useful GPL compliance advice from multiple organizations, in addition to us, over these years.

For our part, since we first contacted Tesla, we have been working with them collaboratively in various ways to convince their original upstream providers, NVIDIA and Parrot, to disclose complete, corresponding source (CCS) releases for all GPL'd binaries found in Tesla's Model S. During that time, Tesla privately provided Conservancy with multiple rounds of “CCS candidates“. (These are source code releases that are not yet complete and corresponding as required by the GPL.) Conservancy in turn reviewed their CCS candidates and provided technical feedback on how to improve the candidates to reach compliance. In this process, we provide detailed reports explaining how the candidate releases fall short of GPL's requirements. This part of the process is the longest, most difficult part of GPL enforcement. We often wish we could celebrate the triumph of moving from a no-source-or-offer violation to the next step of “incomplete sources provided”¹. However, we also can't lose sight of the fact that compliance means meeting all GPL's requirements, so we don't convey false hopes with an incomplete release. We must ultimately remain focused on user freedom in our efforts.

This week, Tesla took a new and different approach. Tesla elected to publish its incomplete CCS candidates, on the online software development collaboration site, GitHub. While our preference is that companies provide adequate CCS immediately, we realize that this can be a challenging process and recognize that Tesla has struggled for years with upstreams to yield proper CCS. We believe Tesla's new approach also has merit, because it allows the entire community to discuss and contribute in public and collaboratively assist Tesla in complying with the GPL. In a case like this, engagement in the community may be an ideal way to transparently assure that compliance is achieved.

We look forward to facilitating Tesla with this new approach to compliance. Toward that end, Conservancy has created a public mailing list to discuss Tesla's source release (and, ideally, to also discuss other CCS candidates if other GPL violators choose to also take this approach.) The first post to this mailing list is our CCS candidate evaluation report ¹, written by our Compliance Engineer, Denver Gingerich.

CCS reports have been the standard document of GPL enforcement since 1998. Conservancy has probably produced hundreds of such reports since we began. However, this marks the first time that circumstances have allowed us to share such a report with the public without violating our Principles. We're excited to do that, thanks to Tesla's willingness to engage everyone in their GPL compliance process.

We know many of you, particularly those Linux-savvy folks who bought Tesla vehicles, have reached high levels of frustration with the lengthy time this GPL compliance effort is taking. Nevertheless, this situation shows precisely why patience is essential for successful enforcement work; it gives us the opportunity to welcome violators to become contributors to the copyleft software community. Our community's history is filled with such success stories. To that end, we ask that everyone join us and our coalition in extending Tesla's time to reach full GPL compliance for Linux and BusyBox, not just for the 30 days provided by following GPLv3's termination provisions, but for at least another six months.

We welcome those interested in the CCS evaluation process to join the mailing list, as this marks one of the few opportunities to engage pubilcly in CCS evaluation. Additionally, anyone who holds copyrights in Linux may join our enforcement coalition of Linux Developers by writing to <linux-services@sfconservancy.org>

---

¹ While Tesla partly corrected the violation yesterday by making some offers for source, the source provided is not complete, corresponding source with complete “scripts used to control compilation and installation of the executable”. Denver's email outlines the specific, current compliance failures.

[permalink]

Tags: GPL

Update on Trademark Action (Fraud Claim Dismissed, New Filing)

by Bradley M. Kuhn on April 30, 2018

As per our commitment to transparency, here's a brief update on the trademark action we have discussed previously.

On Thursday, the Trademark Trial and Appeals Board (TTAB) dismissed SFLC's claim of fraud against Conservancy, as we predicted.

This sensible ruling is just the next step of many; the suit will proceed at the usual near-glacial pace of litigation. On Friday, we moved to the next procedural step, which is to ask the TTAB to allow us to update our answer as planned.

UPDATE: as of 2018-08-24, the Summary Judgement motion is refiled and pending before the TTAB.

We continue steadfast in our previous position: the entire matter remains a waste of resources for both organizations, and SFLC should do the honorable and right thing and simply withdraw their complaint.

[permalink]

Tags: conservancy

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