[OSM-legal-talk] some questions about "Produced Works" under the ODbL
Frederik Ramm
frederik at remote.org
Fri May 21 13:54:51 BST 2010
Hi,
Oliver (skobbler) wrote:
> Consider a "produced work" of something that cannot be reproduced because
> disassembling is not possible. Take a "cake" for example (kitchen wise, just
> for understanding the concept). You cannot produce a another cake from the
> first cake or alter it because you cannot disassemble it to the original
> ingredients since the ingredients have been transformed in a chemical
> process. The cake would be a Produced Work.
>
> If you take a car it can always be disassembled to its individual pieces and
> be rebuild. The car would not be a Produced Work.
That is an interesting concept but I think it is misleading.
What you're saying, basically, is that making a cake loses the (or some)
properties of the ingredients, thus a produced work; making a car
doesn't, thus not produced. (Of course making a car also entails
irreversible actions but let's ignore that for the moment.)
Transferred back to the map database, your argument would be that
anything which does not allow one to go back to the data is a produced work.
However, this is definitely wrong. If I make an excerpt of the full
database, with no information loss, that's certainly not a produced
work. But throwing away some bits - for example by simplifying the
geometries - means that I still have a derived database and not a
produced work, even though the process is not reversible.
The produced work starts where being a database stops. (And yes I know
that's not helpful.)
The proposed community guideline talks of "intent". This hints at the
fact that there may be an overlap; you and I may be producing the exact
same thing from OSM, but yours may be a produced work and mine not.
For example, I could take the bz2'ed planet file and transform it into a
57000x57000 pixel true-colour PNG image with the intent of being able to
reverse engineer that PNG image into the planet file. My image is
clearly a derived database. Now you might be a crazy artist who wants to
give an extra touch to some work of yours, and without much thinking you
take the planet file and slap on an image header and have it displayed.
The image was not created with the intent of extracting the original
data, so it is a produced work - even though the process is entirely
reversible and the original planet file could be created from your image.
On a smaller, and perhaps less unbelievable, scale, someone might create
a special fine-line map rendering from OSM because he wants to
laser-etch the stuff. Produced work. Someone else creates the exact same
fine-line map rendering because he wants to scan it in with his map OCR
program - derived database.
> Also consider the intention behind the concept: The idea is that if someone
> improves the original source then this improvement should go back to the
> community.
This is often said but it is my understanding that, as with GPL, any
improvement you make to the data must only be made available under ODbL
to those who are in receipt of the produced work made from it. This may
or may not be "the community"; it would be entirely legal for me to
improve the data, make a printed atlas, sell you the atlas, hand over
the improved and ODbL-licensed data to you and say: "This is the only
copy of the data. I recommend you shred it right away, as this will make
your atlas unique!"
Interesting question: If you did shred the data, would you be allowed to
publicly display your Atlas afterwards?
Bye
Frederik
More information about the legal-talk
mailing list