[OSM-legal-talk] Produced Work

[Frederik Ramm]

Hi,

SteveC wrote:
> And we'd like help similarly with building a practical definition of  
> Produced Work. Here's how the license RC1 defines it:

Obviously this goes hand in hand with the definition of a (derivative) 
database; everything you make from our data which is not a produced work 
will be a derivative database and vice versa.

> "Produced Work" –  a work (such as an image, audiovisual material,  
> text, or sounds) resulting from using the whole or a Substantial part of the Contents (via a  
> search or other query) from this Database, a Derivative Database, or this Database as part of a  
> Collective Database.
> 
> Thoughts on more practical definitions?

My first impulse has always been to think a produced work (as compared 
to a database) must be like a program binary compared to source code: A 
highly lossy transformation. That would then lead to a definition along 
the lines of  "if the data is sufficiently disfigured it is a produced 
work". But one would always have to tiptoe the line (what about an SVG 
file? and what if the SVG file contains, in comments, every OSM object 
it refers to?)

But someone (forgot who - stand up and claim credit if you read this!) 
has brought up a *very* practical definition during the discussion:

"A produced work is whatever you declare to be one."

It sounds crazy at first, but read on!

This would effectively mean that anyone who makes anything from our data 
has a choice:

* Either declare it to be a derivative database. This means he has to 
put it under ODbL (or a compatible license).

* Or declare it to be a produced work. This means he has a wide range of 
licensing models available, all bound by the reverse engineering clause, 
*and* he has to make the derivative database from which he created the 
work available under ODbL.

In an extreme example, this would mean that someone could take the 
planet file and say: "Here, that's my produced work, and it comes under 
the all-rights-reserved-whatever license." Fine - as soon as someone 
tries to use that planet file by e.g. loading it into a DBMS, the 
reverse engineering clause would kick in: "Creating a Produced Work, and 
then re-creating [...] this Database from the Produced Work, is still 
subject to this Licence."

On the other end of the scale, someone who published PNG tiles on their 
web server could say: "No, this isn't a produced work, we consider these 
  PNG files to be databases." - which would be in line with the EU 
definition of what is a database.

The free choice between produced work and derivative database would 
probably end where the produced work cannot, by any twist of definition, 
be called a database; I guess that would be anything that is neither 
electronic nor arranged systematically in any way, e.g. some work of art 
made from clay or so.

Such a freedom of choice would remove form us the burden to create a 
definition and to watch out for violations. It would of course also be 
attractive for certain use cases; say I created a system that renders 
tiles in real-time and I have applied clever simplification algorithms 
to the data which I would like to keep to myself. It would actually be 
quite attractive for me to say that my published tiles are a database 
and licensed under ODbL, because then that's all I have to release, 
whereas if I declare them to be a produced work, then the license would 
force me to release my cleverly simplified database.

If people are uncomfortable with the full freedom of choice, then we 
could at least create a definition that contains a wide "choice band", 
i.e. the universe of derived works would be divided into (a) some that 
are clearly databases, (b) some that are clearly produced works, and (c) 
a large middle band where people have free choice.

I'm still unsure if this would work, discussion much anticipated, but 
nothing would beat this in terms of practicality!

(Except PD of course.)

Bye
Frederik