[OSM-legal-talk] viral attribution and ODbL

Frederik Ramm frederik at remote.org
Mon Apr 19 20:43:17 BST 2010


TimSC wrote:
> What is the point in paragraph 4.3, if it can be easily side stepped? 

I think it is perfectly ok to state things you would like to see even if 
they can be sidestepped legally. If someone deliberately sidesteps 4.3 
by creating an interim version he releases as PD (but never distributes) 
then he may be in the clear legally but everybody knows there's a bad 
taste to it. That is enough for me at least!

If on the other hand, someone makes something nice from OSM and releases 
it PD, and someone else then makes something based on that guy's work, 
do we really want to require the 38th party down the line to still 
attribute OSM no matter how diluted the OSM content has become? Because 
that's exactly what an attribution chain does.

We have a well working culture of attribution in science, where you 
usually quote the source you took something from, but not the source 
behind the source behind the source.

> And would the large data set import rights holders be happy if they 
> found out? 

Anyone who is not happy with the license need not contribute. (Mind you, 
we're only talking about the "produced work" line here. We do have 
cascading attribution requirements for the body of the data.)

If one wanted to convince those "large data set import rights holders" 
(mind you, some in OSM are of the opinion that we'd better not have any 
imports at all) then one would have to explain to them the following:

OSM and ODbL are about data. Data is what we hold dear, data is what we 
want to have freely available. Data is at the core of what we do. And we 
take measures to guarantee that freedom of the data, including even a 
chain of attribution. The "produced work" is, with that data perspective 
in mind, just a side branch, a by-product, an unimportant dead-end, like 
compiled code. It is not worthy of our special protection; let people 
who make produced works do with them whatever they want.

In the light of this, yes, it can be said that the attribution 
requirement for produced works should be dropped altogether; I think it 
has remained in the license as a symbol. Symbols can be powerful even if 
legally meaningless. "Look, we want you to attribute us, but we freely 
chose not to burden you with tons of license code in order to force you to."

>> We cannot import such data, even today, because we cannot make sure 
>> that the attribution is included in all derived works.

> That seems a bit of a sweeping statement, as far as I understand? Surely 
> our CC-BY-SA license is compatible with CC-BY-SA and CC-BY imports, 
> since the attributions are carried forward? (Although, as Richard 
> pointed out, this is also quite complicated. 

Yes. My interpretation is, somewhat contrary to Richard's, that we are 
in violation of CC-BY-SA and have been from day one, and that thus any 
import under CC-BY-SA or CC-BY is technically not legal. I agree that if 
challenged, Richard's "but what we're doing is reasonable for our media" 
is the best line of defense we can muster.

I am not bothered about individual contributions because everyone who 
contributes *knows* what OSM is like and that he cannot expect to get 
personal attribution. If someone however has released something under 
CC-BY-SA without knowing OSM, they have reason to expect that wherever 
their work is used, they are given credit, and OSM doesn't do that.


Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"

More information about the legal-talk mailing list