[OSM-legal-talk] transitive copyright and "creative errors"

[John Wilbanks]

answering two at once:

jtw

-------
Rob says:

I view OSM as a three-layer stack WRT copyright:

1. At the bottom are the GPS traces or other data. This is basically
factual data, and so not copyrightable. I don't think issues of
personal data or expressiveness come into this.

2. Above this are the nodes, segments and ways, based on that data.
This involves some human decision making, and is a structured data
set. For me, not being a lawyer, this is a grey area and could go
either way.

3. Above this are the maps, rendered from this data, which are
copyrightable. The duration of that copyright is debatable depending
on whether the maps are regarded as computer generated or not.

The nice thing about the ODL is that this stack can be modeled in
licences. But I don't want to see a situation where the "bad guys" can
remove users rights and OSM's licence just helps the good guys stay
good. If the contract approach has that effect, in particular if the
copyleft it creates isn't transitive then I am concerned about it.
-------

I think this is a pretty good summary, actually. The issue is mostly 
about the data in #1 and the stuff in #2. I think #2 could indeed go 
either way - that's probably uncompiled code from a legal perspective 
and I'm not even sure how I would want the courts to decide on it. My 
gut is against the relentless expansion of IPRs, so that leads me to 
lean in the hope that #2 is still PD data, but I understand the desire 
to give people incentives to do the work as well.

But indeed, contracts in the absence of IPRs are non-transitive. That's 
the point of a contract - it's only binding to the people that sign it. 
Otherwise I could sign a contract with Rob and it would bind Steve - and 
something tells me Steve doesn't want me in particular to sign him up 
for anything. We're back to the problem of privity.

When you try to mix a contract-based approach with an open download 
policy, you get into a tough space. Either you have to start a culture 
of tracking and enforcement or you have to expect that if someone - 
maybe not an evil company, maybe just a griefer - reposts your data two 
computers removed from the computer that downloaded it - and that your 
data will then essentially be in the public domain.

------

80n says:

John, perhaps you could comment on the following often suggested gambit:

Interspersed within all the facts in the OSM data dump are a number of
non-factual elements.  There are both accidental errors and probably 
quite a few *deliberate* errors (I know of some).  These non-facts are 
presumably creative works and can be covered by copyright law just like 
any other creative work.  Your suggestion that you can do what you like 
with the data somehow seems to ignore this minor detail.

How would you propose to filter out the non-fact elements that are 
genuinely copyright?

-------

That's a completely new one to me. I need to do some research on it 
before I comment in any detail. But it will pivot on whether or not the 
factual errors are indeed creative works. If they were then in theory 
they could create a risk scenario in which a user would have to assume 
that all the data was potentially contaminated by copyright. But 
wouldn't that also create a distrust in the data compared to data sets 
that didn't have deliberately introduced errors? Questions, questions.

In the life sciences, the better the data set, the closer it is to 
describing the world as it is. And the danger is of unknown errors. But 
every community is different. Let me do some thinking and research on 
this one and I will try to come back with an intelligible answer soon.