Licensing content in Edgeryders: initial thoughts

@fjanss has posed an interesting problem: is it OK to license content on Edgeryders with a CC-BY-SA license, rather than the default CC-BY license specified by the Legals page?

Tentative answer (my own, would be interested in what others think):

  1. There are no ideological problems, of course. CC-BY-SA meets the criteria of the Open Definition (Share Alike is an "Acceptable Condition" to users' rights under section 2.2.3); but even if it did not, it makes sense that each person sets the license of her own project. CC-BY was chosen when Edgeryders was still a Council of Europe project as an even more open license. At the time, we could not envision a spinoff and non-CoE projects riding on the platform.
  2. In fact, it would be simple to add a "License" field to nodes of type project. This is the way GitHub has it: upon creating a repository, you are asked for a license (fun fact: the authors of about 80% of repositories still could not be bothered to choose one). We would not even have to alter the Legals page, that says "Expect where otherwise specified..."
  3. However, this does limit interoperability of within-Edgeryders content. This is only a problem to the extent that we aspire to release open data (see the data strategy for more on this), and this way contribute to the commons. In practice, at the moment (spring 2015) there is only one open data project sitting on top of Edgeryders, and it concerns not content, but data about content. It is a small FLOSS project called Edgesense, a real-time network analysis of the Edgeryders conversation (live dashboard provisional link – code). If you go to the dashboard, you will see a small "download" icon top right: that generates a GEXF graph file. In the final version (almost there now!) this file will be documented and licensed according to the Open Knowledge Foundation's Data Package convention, and pseudonymized. If we had posts and comments licensed under CC-BY-SA or any other license, the license of the secondary data (the graph file, in this case) would break down.
  4. I am seeing three possible solutions to this. One (easiest): restrict the license of the secondary data (graph file) to the most restrictive license used on the primary data (posts and comments). To be able to do open data, the license cannot be any more restrictive than CC-BY-SA. So, in practice, the choice of license on ER would be restricted to CC0, CC-BY, CC0-SA and CC-BY-SA. Two (technically easy, legally clunky): provide for different licenses on the primary data, but ask anyone on Edgeryders to consent to keep secondary data open. Three: limit processability of primary data to those with open licenses (technically complex, legally clean).

@Matthias, are there any flaws in the above?

An interesting problem!

I too, despite being a long time EdgeRyder was actually unaware of this clause in the T&C’s and only came across it when doing the research for this post, where @fjanss first raised this specific issue (linking back because I think there are relevant interrelated issues).  I even checked on waybackmachine to figure out when it was added.

Reading over the open definition you linked to, I realised there is no provision for Creative Commons Non-Commercial licensing. @alberto as this pertains to one of my original questions (about the default ability for the financialisation of content), am I right in understanding from this post that you don’t consider this to ever be an option on ER? Or could you envision the potential for a licensing field that would allow anyone the ability to license as they feel is appropriate to their work? (or at least allows selection from the CC pool of licenses).

I realise this conflicts with some of what you’ve laid out above - but I think it’s possible to encode exceptions.

Given the importance for anyone using and/or contributing to the platform, I think it’s essential that both the T&C’s and the company governance structure are clearly and easily accessible on the site by all users - as noted last week both of these documents seem to have disappeared from view - and on the sign up page they’re slightly obscured in the top right paragraph.

Could you please re-instate my technical admin rights so I can assist with this and other UI issues on EdgeRyders?

Correct

Non-commercial licenses are not considered open – we had a long and interesting debate on the matter around 2010 in the open data community. It came down (for developers) to ensuring interoperability, and (for lawyers) to focus on enabling, rather than restricting, users. That does not mean it should not be considered as an option: we could decide to enable non-commercial licenses (or, on the contrary, full copyright). There is a price to pay: this rules out my solution 1, and additionally – I fear – is very difficult to enforce in the long run.

Please understand that this is more about sharing a space than it is about legals. Nobody can stop any Edgeryders user from writing something like:

The sentence in red in this post is licensed under a GNU General Public License.

This is legally valid: I have just broken interoperability on the platform. However, in practice, the platform does not know that: it is up to social interaction and negotiation to enforce that license. Example: @mariabyck objected to have her posts contribute to the ethnography of stewardship we are preparing for the Rockefeller Foundation. That was solvable and solved, but it required a specific conversation. In two years, researchers from the University of Nevada or something could be scouring the web and doing ethnography on her posts, and it would be legal to do so under the terms of the license. My personal opinion is: it is probably simpler and safer that, if you do not mean to release open content, you do not use Edgeryders at all for it.

I suggest you speak with @Matthias, who is clearest on the hidden long-run difficulties of maintaining solutions and not just building them… as I was writing this comment I see he has come online.

It is definitely a good idea to make legals accessible from every page. I have put up a provisional hack (link “Legals” in the main menu, referring to a much more complete page that @Nadia (thanks!) has put together a couple of weeks ago. It was intended for the company website (and it has that look-and-feel) but it does contain all the relevant information. For a more elegant solution, I suggest you, once again, speak to Matt. We have looked into several possibilities (footers, two-level menus…) but all of them have pitfalls.

Also: interoperability as contribution

By the way, I just realized that the example above is also great in terms of what happens downstream of the licensing. Since Maria preferred it that way, Inga did not include her stewardship case studies in her ethnography. That ethnography is part of a relationship between ER and Rockefeller Foundation, in which RF paid for some of the costs of the LOTE4, which came as free and even with a few goodies (a stocked kitchen, travel grant support etc.) to the community. So:

  • The community discussed about stewardship before, during and after LOTE4, and licensed that discussion as open content.
  • ER distilled an ethnography out of that discussion (the final report is being written now, it will of course be shared back and in turn made available as open content).
  • The ethnography was supported by RF, so that RF ended up supporting LOTE4 in a quite direct way.

The result of all this is that a discussion on stewardship, that the community wanted to do anyway, ended up generating some support for the people who were having it! This is exactly the kind of stuff I think ER should be doing. It was only possible with open content. With non-open content, the ethnography would be illegal and hence unfundable. So, allowing your content to be open and interoperable with others’ is a very direct way to support your work and that of others. There was no problem with Maria’s opt-out – there was plenty of content to go around – but if everyone had opted out it might have been impossible to deliver LOTE4 as we experienced it.

Thank you!

The clarity and time to explain is appreciated.

When I raised the question with regard to NC, this was actually one instance I had in mind - and wasn’t aware that Maria had requested this - are there links to that discussion? Or was it offline at LOTE4?

Just to further understand, does this then literally apply to every piece of material on ER? - e.g user profiles, user avatars, the admin forum and other closed groups?

Not sure

I am not sure, Ben. @Noemi will know.

As for pieces of material, again you are correct. “Except where otherwise noted, content on this site is licensed under a  Creative Commons Attribution 3.0 Unported License…” [source]. It also makes sense from an interoperability perspective: when the server gives you a page, it is remixing material from different repositories and in different formats, including text, images and markup. If the licenses on the different components were different, it would become a nightmare to figure out who can see what. So we just stick the same default license on everything.

Maria voiced concerns online and offline

@Ben, it was a mix: we had the Open Ethnographer, ER LbG transparency sessions at Lote4 and she did not seem convinced, nor after the meeting afterwards. She voiced her concerns partly on the platform and then on email, concluding between the two of us.

I can’t speak for her, but for me it was important to have acknowledged the good faith and transparency of the company: the Rockefeller contract came in after we decided to have Lote4 and the case study adventure, and after everyone had it clear that we can only run Lote4 with all contributing - some with content, some with logistics, some with engagement etc. So there were many types of work going in, some more visible or subject to enforce licensing than others. Anyway, hopefully we all learned from that.

Copyleft or permissive

Quoting Alberto: “If we had posts and comments licensed under CC-BY-SA or any other license, the license of the secondary data (the graph file, in this case) would break down.

In the case of Edgesense, there are no verbatim takeovers of content from edgeryders.eu posts and comments, and only these are covered by content licencing. Taking over facts does not infringe any licence (exception: in the case of database records, for countries having extended copyright to databases; but that’s not about pure metadata as here). So, no, I don’t think CC-BY-SA licencing of content would have any effects on Edgesense secondary data. It would have effect on Open Ethnographer’s secondary data. But if a copyleft “share alike” clause is the only “restriction”, no issue at all. Personally I like both the permissive and copyleft style open licences :slight_smile:

On a “non-commercial” clause: it is kinda agreed in the software world that “free” licencing means not infringing against a use, so a “non-commercial” clause means that the licence is no longer open. Accepting unintended use is at the core of being “open”: both the price one pays for opening ones stuff, and the benefit one gets by being allowed to do everything whatsoever with others’ open stuff (if attributing and potentially sharing on, granted). So, I will be a Stallman here … I’m against non-commercial clauses, IMHO they only do harm in any ecosystem wanting to operate by being open. And besides, I so far could not figure out what they mean: preventing from selling the licenced work, or preventing from using it internally in a company as well?

“In the case of Edgesense, there are no verbatim takeovers of content from edgeryders.eu posts and comments, and only these are covered by content licencing. Taking over facts does not infringe any licence (exception: in the case of database records, for countries having extended copyright to databases; but that’s not about pure metadata as here). So, no, I don’t think CC-BY-SA licencing of content would have any effects on Edgesense secondary data. It would have effect on Open Ethnographer’s secondary data.”

This makes more sense to me now. I was getting confused by the different licensing terms that were being applied to the different projects. I didn’t realise that the Edgesense project was under different terms to the Edgeryders.eu website content.


“But if a copyleft “share alike” clause is the only “restriction”, no issue at all. Personally I like both the permissive and copyleft style open licences :)”

Why would we not use the GPL? I can see why you would want to use some form of a Creative Commons license, but in your opinion, what are the reasons against using the GPL?


"On a “non-commercial” clause: it is kinda agreed in the software world that “free” licencing means not infringing against a use, so a “non-commercial” clause means that the licence is no longer open. "

I’ve only found this to be the case in the corporate sponsored FOSS software, as opposed to the FLOSS software. The older projects that started before Open-Source became another marketing buzzword used one of the GPL variants, because that was what was available then. When companies wanted to contribute to those projects, they had no choice but use the GPL for thier contributions.

There were a number of commercial software companies, where they had usage restrictions as part of the licensing. It was a way of reducing and removing liability. You still can’t use any MicroSoft software to control any form of nuclear system… :))

There were other Open-Source Software licenses, that were in use before the GPL was developed, which had similar restrictions when it came to safety-critical systems. While this was a way of removing liability from the developers, it also was a method of empowering the users with idea of being personally responsible for the software that they were running, as well as providing them with the capability of being able to do so, without being tied to a specific vendor.

The mainframe market has used this method since the 1960’s, and it’s still where the enterprise market makes a lot of their margins.


“Accepting unintended use is at the core of being “open”: both the price one pays for opening ones stuff, and the benefit one gets by being allowed to do everything whatsoever with others’ open stuff (if attributing and potentially sharing on, granted).”

When it comes to tool making, this is one of the tool-maker’s main protection from being sued and/or shutdown. You can use the same CNC system to make food-growing devices, or you can use it to make artillery.

To mis-quote @Hexayurt, because the price-point of automated manufacturing is sliding down the scale from full-scale industrial processes, to the personal scale, it changes the way we have to do things…


“And besides, I so far could not figure out what they mean: preventing from selling the licenced work, or preventing from using it internally in a company as well?”

This i have never been able to figure out.

This sort of restrictions almost always went hand-in-hand with the programmers email address, so if you wanted to discuss dual licencing for a specific use, you could find them, and negotiate a direct contract. There are two hardware programmer’s i know who were doing this back in the early 90’s.

I couldn’t get a straight answer from them then either… :))

It’s a question that i need answered, in a legally-watertight manner, before i start releasing any of the tools i’ve designed.

I’m not against people making and USING the tool designs for themselves. I just want to make sure whenever they make and SELL a tool that’s made from an OSHW design, the person who is receiving the machine, also receives a copy of the design docs as well, so they are not tied to that specific vendor.

I also want to work out some clauses, so that if the machine is made and HIRED out, then the person hiring the machine gets granted the same access rights to the original design documentation, under an identical set of copyright restrictions. And this will help fix another trap that Stallman foresaw.


I remember some of the FUD that was being used by Microsoft around the back end of the 90’s, where they were saying that every document created with a GPL word-processor would automatically be given a GPL copyright license. It was basically another attempt at confusing middle-managers.

However, you can see where Stallman was coming from, when you realise that when the earliest compilers were running on mainframes, they had a usage licence where the copyrights of the compiled software was set by the company that made the compiler.

His hack was to make a second compiler that stuck to those restrictions, but didn’t have any restrictions on the output of the second compiler. The first thing he did with the second compiler he had made, was to compile the third version of the compiler, which was free of any restrictions whatsoever. And that’s what turned into the GCC. :smiley:


Maybe you need patent / design pattern protection, maybe not …

Love your tale of the GCC history, didn’t know it.

in your opinion, what are the reasons against using the GPL?

For software, no reasons against it. For website content? Well it “works”, but c’m on, it’s ugly and confusing to make people read a licence talking about linking libraries and stuff when it’s about textual content. Or wait, I know how to explain it, Mr. toolmaker: eyery tool for its purpose! :slight_smile:

It’s a question that i need answered, in a legally-watertight manner, before i start releasing any of the tools i’ve designed. I’m not against people making and USING the tool designs for themselves. I just want to make sure whenever they make and SELL a tool that’s made from an OSHW design, the person who is receiving the machine, also receives a copy of the design docs as well, so they are not tied to that specific vendor.

Interesting problem! Quick thought: When basing your licence on copyright laws (which is easiest and does not cost patent application fees and stuff), you can only meaningfully have the rules you mention for digitally manufactured parts. For anything that requires users to download a file with data (3D geometry, gcode, whatever) to their computer. Because when they make a copy, copyright applies, and you can put in the conditions you like into the licence grant. Just like GPL requires shipping the code with the binary (or on request) – and the binary would be based on your code. However, if users can make a tool from OSHW documentation without making local copies of the documentation, you have no legal say via copyright. And viewing something in the browser does not count as making a local copy (in Germany at least, and I guess in many other legislations).

“For software, no reasons against it. For website content? Well it “works”, but c’m on, it’s ugly and confusing to make people read a licence talking about linking libraries and stuff when it’s about textual content. Or wait, I know how to explain it, Mr. toolmaker: eyery tool for its purpose! :)”

And simple tools that each do one thing well, with clear interfaces for interoperability. So far, so Unix. :smiley:


Copyright protection only goes so far with hardware. It’s really specific to the design docs, and not the objects made using the design docs. You could restrict the distribution of the made objects without a design docs distribution, by having copyright licenses to the design docs using clauses that would mean if they didn’t distribute the docs, then they’d be guilty of copyright infringement, as they’d be using the original docs without a valid license.

It won’t stop anyone from reverse-engineering the made objects, or just copying the designs and using them without attribution, but it’s another speed-bump for bad actors, and another carrot for good actors.


Patents are a whole 'nother issue. Each patent contains specific claims that will detail how something is done, but it will only cover the method that is explicitly laid out in those claims. It doesn’t cover any work-arounds, or alternative methods for doing the thing that is specified in the patent.

They are also country-specific, so you’ll need to file a patent application in every country at once. Is why most companies try for the USA patent system, as they’re vastly under-funded, under-staffed, and the resulting patent will cover a wide geographic area, that is in theory one unified market.

Improving the market disparity between the EU and the USA, is one of the strands of thought that is behind the recent (ie. last 15 years.) attempts to harmonise all of the separate patent systems that you find in the different countries in the EU. Though as you might guess, there are already bad actors trying to game the system, even as they’re building it.


The other thing about patents is that they only give the right to sue, nothing else. Yes, you can get injunctions that will bar the importation of the physical objects, and, yes, the courts are willing to serve those, but you have to be able to afford to go to the courts in the first place. The patent courts are a rich person’s game.

And time that is spent suing people is time that you aren’t using to make the next iteration…

Don’t get me wrong, i like the level of disclosure that is involved in the patent system, and the relative level of protection that they provide. A paper shield is better than none, but it requires a large wallet to make any meaningful attempt to enforce the patent.

You also get the problem that sometimes the big cartels will shut a patent-holder out of the market, untill the patent only has a few years left to run, and then give a take-it-or-we’ll-wait-till-it-expires offer to the patent-holder.

And yes, that sort of market-manipulation IS illegal, but for some reason it doesn’t get prosecuted very often, if at all…

A nice example of this, was the invention of the recordable digital disc. Invented and patented in 1964, but only released to the market around the early '80’s, approximately 2 years before the expiry date…


And you’ve already seen where i’m eventually heading. Boot-strappable CNC systems, that you can make the initial versions with hand tools and hand power-tools, that you can use to upgrade themselves to full functionality. (And yes this was partly inspired by the GCC story, partially by the Reprap project, and partially by David Gingery.)


The legal protection that i’m looking for will be ignored by bad actors, who’ll just take the design docs and use them, while ignoring the licences. Have you tried to sell technology to China recently? They’ll only buy what they can’t copy or reverse-engineer, and that’s even before you start to talk about the intermittent enforcement of the Berne Convention by the Chinese courts.

By using only technology from expired patents, you sidestep almost all of the boobytraps that are a direct result of the patent systems, and you are guaranteed to get the design’s that have been thoroughly de-bugged. Very important when you are designing safety-critical systems, and all forms of infrastructure are safety-critical systems.


I’ve been researching the use of the EU Registered Design and the licencing and sub-licencing of those designs. It’s possible to create a GPL-like distribution licence, however it does mean that you’ll still have to deal with a gatekeeper bureaucracy that decides whether it’s a valid design…

Maybe i’ll just have to use Vinay’s methods and hire some ghurka mercenaries… :))

“Simple” answer

To try a simple answer: There is absolutely nothing against changing the terms so that users can select from various open licences for their content. Selecting would be on project level, or if we want to really go all the way, through a field on node / comment level.

The solution for secondary data that includes quotes (not a problem for Edgesense!) would be Alberto’s first “easiest” alternative: restrict the license of the secondary data (graph file) to the most restrictive license used on the primary data (posts and comments). Or if we want, mark up all verbatim takeovers in the secondary data with their respective licence, as indicated by users originally. Having secondary data under copyleft (with a “Share Alike” clause) is preferable anyway because it keeps research ethical …

For interoperability and mutual benefits, I’d suggest that only open licences (incl. ShareAlike ones of course) would be available as choices when posting content. To use an analogy, edgeryders.eu is like a large software code repository.

@Ben: a CC-BY-NC-* licence does not protect against somebody taking an idea or name and monetizing it. Just against somebody taking an exact copy of a text, logo, work of art etc. and monetizing it. Redrawing the same logo from memory and selling that is not prevented by CC-BY-NC-*. Since viewing public web content does not require signing a T&C document, I think the rules you’re looking for can’t be had at the same time when as when having public content. Except when going via the expensive routes of trademark protection before publishing a name / brand, or via the patent route before publishing a technical invention. I don’t think there is a legal route how concepts and non-technical inventions can be protected, except through secrecy and the T&Cs of a closed online space. But anyway, this is another interesting discussion, beyond the content licencing. Let’s have it in time!

Thanks + protecting concepts

Thanks, for my present project all proposed variants are ok, and probably the simplest is the best.

Concerning the problem of protecting concepts. As @Matthias says it is another discussion, but just for the record : as I alluded to in A proposal for testing Assembl I do not think it is possible (for us in the present circumstances) if they are perceived as valuable for ‘serious’ monetarised exploitation, and that is why I started my present project.

As a first approximation the protection you can have is only as strong as the budget of your legal department, compared to that of your potential competitors.

So, time for wrapup

This concerns immediately @fjanss and @Matthias: so what’s next? Do we actually go on and implement a license field on nodes? At what level? Actually it might be technically easier to attach one to each and every unit of content, down to the comment level (no complicated SQL query to inherit the license from the node or group node). We set the default to CC-BY, and it’s done.

And: do we just admit open licenses, with a drop-down widget? Or do we allow NC and full copyright? I think we should stay with open licenses, for exactly the reason that Matt and Fjanss mention: licenses do not prevent people re-using you concept; and anyway your legal protection is a function of the budget of your legal department. Better to focus on enabling reuse.

level of licence setting

If it is easy to do, the best would be to be able to change the default lincence at the project level. Otherwise it would be a chore to remember changing the licence at each unit of content. And it would enable fine grained settings for the case it is needed.

The idea of a project owner being able to set this when the project is being created, will be incredibly helpful.

You’ll need to add in a time-tracking system of when changes are made, and who was making the changes. Having that sort of logging data backed up in a publically available place will solve most arguments before they start.

This is entirely the sort of things that webservers and time-based automatic scripting (a Cron-Job) was designed for. It means that we’ve already got the tools. We just need to build the infrastructure to make good use of them. :smiley:

Being able to use finer-grained settings for each module, would mean that you’d be able to take advantage of modules that were released under different licences, as long as you’re using only the boundary interfaces, (In computer terms, this is called the linking exception.).

And this in turn will allow the use of systems that you hire, or buy in, whilst working on building your own versions.

Please add in a transparent system of access controls, as to who can make changes, otherwise it’ll just be another implementation of Orwell’s Memory Hole.

This looks really hard :frowning:

But Matthias has last word, as the guy who has the ultimate responsibility for keeping the website up.

The only thing that I see easy is ensuring user control over licensing: every user can already edit her own content. “Edit” means all fields can be modified. If we add a license field, it will be editable by the author of each unti of content, just like any other field. Keeping track of variations in time looks more difficult… and yet Billy is right, it would be useful. I have no idea how to do it. The cron job seems really clunky, one more piece of software to keep alive… sad

Field revisions

Drupal has field revisions, this would be the way to go to keep track of licence changes. The alternative would be to only allow changing the licence to something more permissive when editing existing content (like, CC-BY-SA to CC-BY to CC0). The other direction makes no real sense anyway since a user could always take the old, more permissively licenced version from waybackmachine etc. and use that. (Well, that other direction makes sense when substantial edits are made, but these are very rare. Always design a feature for the 90%!)

Another interesting case will be the licence field for wikis. I think it should be non-changeable after creating the wiki. Else, every contributor would have to agree on the licence change, and we don’t want to implement such a feature that would only get used 0.5 times over its lifetime :slight_smile:

I just see Drupal modules creativecommons and creative_commons. Sweet! Both provide CCK fields, and allow defaults. The part missing is only the “default per group” setting.

choice of module

@Matthias, have you made a choice of which module to start with ?