There is lots of legal chat around WMF, and there are lawyers involved in the projects. Remarkably, they are very hands-off, preferring to take the path of openness, fairness, least harm to individuals, and most benefit to humanity.
Here are some things I heard about legal issues...
Handy casebook of examples
UMich Casebook — This Is Not Legal Advice! (But it is awesome.)
Geoff Brigham, WMF Chief Counsel
Greg Grossmeier, Creative Commons
Greg gave some clarification to some aspects of CC licenses:
- ND or No derivatives is defined in local law
- CC0 waives all rights to the extent that this is possible in local law (e.g. you cannot relinquish your moral rights as author in France)
- Public Domain is not the same as CC0
- NC or Non-Commercial — "whatever the hell that means"
- BY vs BY-SA is "a matter of personal choice"
He said that they recommend CC0 for scientific data. I don't like this much, because I think attribution is central to reproducibility in science. I prefer CC-BY, which gets out of the way. However, I have heard some people saying that CC is not suitable for data, since many types of data are not really creative and therefore not covered by copyright, and therefore not easily connected with Creative Commons.
Deror Avi on images in Commons
In the US, you cannot make a copyright image of a non-copyrighted item. For example, you can't copyright a photo of the Mona Lisa. You can do this in the UK, however (I recently paid over GBP100 to use a photograph of a page from Leonardo's early 16th century notebooks).
The so-called Freedom of Panorama applies to images of buildings and other public structures, but also depends on the country:
- In most countries, it includes all public buildings and sculptures — photos of these things are copyright-free
- Only covers buildings in the US, not sculptures
- Includes all public art in Israel, especially interesting as some works appear in the US and in Israel
- In Italy, the is no Freedom of Panorama at all!
There is a de minimis exception, however. If the object is not the subject of the image, but only a small part of it, then it's OK — the image can be licensed under CC terms.
All this is especially interesting because the Commons servers are in the US. Most of this is yet to find precedent in case law.
Random observations about open data
- It seems most people think Creative Commons is not a good fit for licensing data. This is because copyright only covers creative works, and data — depending on your interpretation — is not always 'created' in the same way as, say, a book. You could probably argue the case for seismic data and even earth models, which are creative in ways that the census, say, is not (or not intended to be!). OSR uses CC-BY-SA , so there's precedent there, but no legal precedent (I suspect).
- Open Street Map is moving from CC to the new open database license (ODBL) from the Open Knowledge Foundation. I don't know much about this license, and this was the first I'd heard of it.
- Some countries have introduced something called 'database right' for data. This is intended to act like copyright, but recognize the lower level of creativity in assembling a database. I don't think CC fits with Database right, but maybe ODBL does. I don't know; this was the first I'd heard of it.
- The thing about copyright is that you don't have to assert it. It's just yours, when you create something. However, to claim damages for copyright infringement later, you need to register it with the Copyright Office (or equivalent). Most people don't do this. I don't know if it's the same for Database right.
- Some people recommend CC0 (essentially public domain, though there is a difference—see above) for data. I don't like this myself, because it relinquishes the right (and obligation of) attribution, which is an important part of reproducibility. I think we need the BY clause, or something like it, in scientific licensing.
- Most people also recommend against CC for software. I recall Victoria Stodden mentioning this at EAGE. It seems the way to go here is MIT- or BSD-style permissive licenses, or the GPL for non-permissive rights. Another option for libraries, as opposed to applications, is the LGPL, which is not as aggressively 'open' as the GPL. I use the modified BSD license myself.
- If I had to choose something today, I'd go for CC-BY (permissive) or CC-BY-SA (non-permissive) — one can always change it later, as OSM are showing. Wikipedia also changed its license several years ago (from GFDL to CC-BY-SA). Of course, it's hard to make it less permissive later, because by then it's 'out there'.