There is no Free Software company - But!

Since the start of the FSFE 15 years ago, the people involved were certain that companies are a crucial part to reach our goal of software freedom. For many years we have explained to companies – IT as well as non-IT – what benefits they have from Free Software. We encourage individuals and companies to pay for Free Software, as much as we encourage companies to use Free Software in their offers.

While more people demanded Free Software, we also saw more companies claiming something is Free Software or Open Source Software although it is not. This behaviour – also called "openwashing" is nothing special for Free Software, some companies also claim something is "organic" or "fair-trade" although it is not. As the attempts to get a trademark for "Open Source" failed, it is difficult to legally prevent companies from calling something "Free Software" or "Open Source Software" although it does neither comply with the Free Software definition by the Free Software Foundation nor with the Open Source definition by the Open Source Initiative.

When the FSFE was founded in 2001 there was already the idea to encourage and support companies making money with Free Software by starting a "GNU business network". One of the stumbling blocks for that was always the definition of a Free Software company. It cannot just be the usage of Free Software or the contribution to Free Software, but also needs to include what rights they are offering their customers. Another factor was whether the revenue stream is tied to proprietary licensing conditions. Would we also allow a small revenue from proprietary software, and how high is that that you can still consider it a Free Software company?

It turned out to be a very complicated issue, and although we were regularly discussing it we did not have an idea how to approach the problems in defining a Free Software company.

During our last meeting of the FSFE's General Assembly we came to the conclusion that there was a flaw in our thinking and that it does not make sense to think about "Free Software companies". In hindsight it might look obvious, but for me the discussion was an eye opener, and I have the feeling that was a huge step for software freedom.

As a side note: When we have the official general assembly of the FSFE we always use this opportunity to have more discussions during the days before or after. Sometimes they focus on internal topics, organisational changes, but often there is brainstorming about the "hot topics of software freedom" and where the FSFE has to engage in the long run. At this year's meeting, from 7 to 9 October, inspired by Georg Greve's and Nicolas Dietrich's input, we spent the whole Saturday thinking about the long term challenges for software freedom with the focus on the private sector.

We talked about the challenges of software freedom presented by economies of scale, networking effects, investment preference, and users making convenience and price based decisions over values – even when they declare themselves value conscious.

One problem preventing a wider spread of software freedom identified there was that Free Software is being undermined by companies that abuse the positive brand recognition of Free Software / Open Source by "openwashing" themselves. Sometimes they offer products that do not even have a Free Software version. This penalises companies and groups that aim to work within the principles of Free Software and damages the recognition of Free Software / Open Source in the market. The consequence is reduced confidence in Free Software, fewer developers working on it, fewer companies providing it, and less Free Software being written in favour of proprietary models.

In the discussion, one question kept arising. Is an activity that is good for Free Software which is done by one small company as their sole activity more valuable than if the same thing were done as part of a larger enterprise? We all agree that a small company which is using and distributing exclusively Free Software, and has done so for many years, and no part of the software they wrote or included was ever non-free software is good. But what happens if said small, focused company got purchased by a larger entity? Does that invalidate the benefit of what is being done?

We concluded that good action remains good action, and that the FSFE should encourage good actions. So instead of focusing on the company as such we should focus on the activity itself; we should think about "Free Software business activities", "Free Software business offers", and such. My feeling was that this was the moment the penny had dropped, while others and me realised the flaw in our previous thinking. We need action oriented approaches and we need to look at activities individually.

There was still the question where to draw the line between acceptable or useful activities and harmful ones. This is not a black and white issue, and when assessing the impact for software freedom there are different levels. For example if you evaluate a sharing platform, you might find out that the core is Free Software, but the sharing module itself is proprietary. This is a bad offer if you want to run a competing sharing platform using Free Software.

The counter example of an acceptable offer was a collaboration software that was useful and complete, but where connecting a proprietary client would itself require a proprietary connector. It was also discussed that sometimes you need to interface with proprietary systems through proprietary libraries that do not allow connecting with Free Software unless one were to first replace the entire API/library itself.

Ultimately a consensus emerged around a focus on the four freedoms of Free Software in relation to the question of whether the software is sufficiently complete and useful to run a competing business.

One thought was to run "test cases" to evaluate how good an offer is on the Free Software scale. Something like a regular bulletin about best and worst practice. We could look at a business activities and study it according to the criteria below, evaluate it, making that evaluation and its conclusions public. That way we can help to build customer awareness about software freedom. Here is a first idea for a scale:

• EXCELLENT: Free Software only and on all levels, no exceptions.

• GOOD: Free Software as a complete, useful, and fully supportable product. Support available for Free Software version.

• ACCEPTABLE: Proprietary interfaces to proprietary systems and applications, especially complex systems that require complex APIs/libraries/SDKs, as long as the above is still met.

• BAD: Essential / important functionality only available proprietary, critical functionality missing from Free Software (one example for an essential functionality was LDAP connector).

• REALLY BAD: Fully proprietary, but claiming to be Free Software / Open Source Software.

Now I would like to know from you: what is your first reaction on this? Would you like to add something? Do you have ideas what should be included in a checklist for such a test? Would you be interested to help us to evaluate how good some offers are on such a scale?

To summarise, I believe it was a mistake to think about businesses as a whole before and that if we want to take the next big steps we should think about Free Software business offers / activities – at least until we have a better name for what I described above. We should help companies that they are not deluded by people just claiming something is Free Software, but give them the tools to check themselves.

PS: Thank you very much to the participants at the FSFE meeting, especially Georg Greve for pushing this topic and internally summarising our discussion, and Mirko Böhm who's contribution was the trigger in the discussion.