Interpreting the Open Database License
For reference, here is the ODbL. There is a nice human readable summary. You can also read more in the Wikipedia entry.
The most famous database available under the ODbL is OpenStreetMaps.
I recently found out about OpenCorporates, which is a global database of companies, published under the ODbL. I thought this was great, so I applied for access to use the database for a project. I was denied because I'm not a journalist or a nonprofit and instead was invited to pay for access instead. And it's not cheap, likely because company databases are often used in the B2B space.
I replied that this seemed to be in conflict with their mission, especially given that my project was focused on using the data to create a benefit to the public, and their response was that they wanted to protect against their database being copied.
From my reading, this seems to be in direct conflict with the ODbL. Egregiously so, which has me thinking I'm missing something.
Does anyone have any insight? It seems to me that the whole point of the ODbL license is to make data freely available. This is backed up by interpretations I came across while searching and by the ethos of other orgs using the license, such as OSM. What am I missing?
Edit: I'm still excited to hear from anyone with knowledge in this area, or just general insights into how I'm misunderstanding the license.
And also, having learned that The Open Data Commons, which publishes and maintains the ODbL, uses this definition of the concept of open... I'm leaning towards the interpretation that OpenCorporates wants the aura of using a reputable license with the word "open" in it, but isn't genuinely interested in the ethos. Which is disappointing but not shocking, they'd be far from the first.
If it's published under ODbL but they don't give access openly then you could find it from someone else. Them licensing it a certain way doesn't compel them to always be able to physically distribute it freely:
https://opendata.stackexchange.com/questions/15898/why-is-opencorporates-considered-open-data-when-nothing-can-be-downloaded
Even if they gave you the free API key, it seems severely limited to the point of being useless:
Thanks for the link. One comment there links to a version of the definition of "open" as it pertains to the license. It would seem that they fail under that definition.
Maybe they consider themselves to be using an earlier version of the license where the definition of open is less explicit. Or maybe they just don't care about the license and are yet another company using the word open, and an open license, cynically for social cred points.
That occured to me, but I'm not sure where I would find another source for the database. Any registered journalists on Tildes?
Regarding the API, yes the restrictions are pretty comical, but they also offer, or claim to offer, a full download of the database to qualifying orgs.
I think this specifically is the wrong conclusion--you were right before. Their use of ODbL is wrong because they treat it as a content license but the purpose of the ODbL specifically relates to database rights.
Actually, they even disclaim certain content rights because much of the content is public domain. While they have certain database rights due to the effort they exerted to file FOIA, data processing, etc--but at the same time make it as difficult (or perhaps even moreso difficult) for you to compile the same data (from them or due to their prior actions)--in some sense their use of ODbL on the small subset of data that they give you doesn't actually apply at all--unless you got the full or substantial portion of the database from them or someone else under that license.
Facts are public domain. That doesn't mean that they are required to distribute them freely--they can hold the data like a trade secret--but that behavior doesn't really sound like the spirit of free and openness! They should put the data on BigQuery or S3 requester pays.
But all this contemplating doesn't really matter because judges, lawyers, and laws often contradict themselves and it often seems the case that whoever has the pointier stick, wins--and, depending on your country, ODbL may not even have any legal standing (other than via contract law). Sui generis Database Rights are only recognized in a small number of countries.
It sounds like our reading of the ODbL is a bit different. I think you're right that it's meant to be a license that considers rights relevant to the database as potentially distinct from content rights, but it also has language that specifically covers content (for example: copying, modifying, distribution and sharing). In their FAQ it talks about a use case where the rightsholder for the content is the same as for the database, or where there are no relevent rights to the content (like public domain). In that example the content can be covered by the ODbL and no other license is necessary. In other cases is suggests a secondary license which, as far as I can tell, OpenCorporates doesn't use.
I also wonder how important the distinction between data and content is in this context. The data in the database and the contents of the database are essentially the same thing as long as they don't have conflicting rightsholders. In that context content might instead refer to something produced from the result of a database query (like a pie chart).
In any case my reading is that the ODbL can be used to refer specifically or solely to database rights but it can also be used as a license for both the database and its content.
After that it comes down to the definition of open. The Open Data Commons, which publishes and maintains the ODbL, uses this definition.
Wow! This is very relevant to a personal project I've wanted to do for years. I'm curious what your use case is?
I've long wanted to create an app/website that uses corporate relationships + manually curated ethical profiles to help people find products from companies aligned with their values. Not an affiliate/profit driven product, but a non profit designed to simply solve the problem of helping consumers find businesses based on value alignment.
This is a digression, but your post gives me something to think about. I'd abandoned the idea because of the effort of maintaining the corporate connections, but maybe this could be helpful.
If only it was as open as it first appears. If you figure out how to get access to the data, please let me know!