[OSM-legal-talk] clarifying

[John Wilbanks]

I'm about to head onto a plane for the next day or so, but I wanted to 
address something I sensed in the last email exchange.

Hopefully if you're interested in what I'm typing here, you've actually 
read the protocol on OA to data - it's at 
http://sciencecommons.org/projects/publishing/open-access-data-protocol/ 
and it's by far the clearest explication of my position here. Please 
read it if you've not done so, at least, if you want to go back and 
forth with me, as it lays out the end of a long research project. A few 
points to call out from it, for this argument. All of these are in the 
first person because they represent my own *opinions* and not statements 
of fact.

- I am not trying to say that databases don't have copyright - I'm 
saying that about data, not databases. I have been trying to articulate 
where the general consensus about copyright stands (thus, when there is 
a copyright on databases, it comes from arrangement, not easter eggs). 
But it's very clear to everyone that there is copyright somewhere in 
DBs. That's in the protocol.

- I am trying to say that the danger of using the ODL approach versus 
the PDDL approach (which is the one I prefer - Jordan's done both of the 
big ones!) is in the *propagation* of asserted rights from the database 
down to the data, because it's so hard to figure out where copyright 
stops and starts. Viz our debate here.  This is also laid out in more 
detail in the protocol - the risk that users will assume the copyright 
extends to the traces in this case from the maps.

- I believe the risk of this propagation of asserted rights poses 
long-term problems for the utility of data. To make a simple example, I 
am in discussions with some holders of clinical trial data about mixing 
it up with geospatial data, so that responses to trials and cancer 
prevalence patterns can be mapped by neighborhood. Clinical trial data 
is subject to privacy requirements which are totally incompatible with 
the types of contractual assertions made in the ODL. Thus, OSM is not a 
candidate for use in this mashup project.

- I believe that this type of data mashup approach is going to increase 
in variety, complexity, and importance over the coming yeas. I also 
believe that the public domain approach, in which the owner *knows it 
has some rights in the database but waives them*, is the approach that 
makes a given database more likely to be a foundational source than others.

That is the freedom that I am arguing for - the freedom to integrate 
this data into other kinds of data entirely, which may well be totally 
incompatible with share-alike for lots of reasons, or which we don't 
even know about yet. The share-alike contingent of this list argues for 
a different kind of freedom, in which the provision of your data to them 
requires the recontribution of data. Two different freedoms, competing 
with each other...it is that distinction in desired freedoms that forms 
that basis of our disagreement, and it's a legitimate distinction for us 
both to make.

jtw