[OSM-legal-talk] CT clarification: third-party sources

[David Groom]

> ----- Original Message ----- 
> From: "Frederik Ramm" <frederik at remote.org>
> To: "Licensing and other legal discussions." 
> <legal-talk at openstreetmap.org>
> Sent: Thursday, December 09, 2010 7:14 PM
> Subject: Re: [OSM-legal-talk] CT clarification: third-party sources
>
>
>
> Peter,
>
> pecisk at gmail.com wrote:
>> 1) I'm not against ODbL. It is nice idea and I wholeheartedly support it;
>> 2) I'm not against general idea of CT, I understand why it is needed;
>>
>> My confusion and problem lies within fact, that while I can accept CT
>> if I add only my own data to OSM, I can't to do that due of
>> third-party sources because some of them requires attribution and
>> share alike.
>
> Do you have a concrete example of a third-party source that does not
> specify a concrete license, but requests a general "attribution and
> share alike"? Or is this only theoretical?
>
>> While ODbL is good enough for both of these things
>> (theoretically), then CT blocks, because it says that nature of the
>> license of imported data can change. As I'm not author of those data,
>> I don't have permission to change nature of the license.
>
> In the original setup, data that is not compatible with the CT would not
> be accepted, so if someone required attribution and share-alike, that
> data would not be compatible.
>
> The current mood in LWG (as per the latest CT draft) seems to be to

Whilst a "mood" might to some people be enough, it would be fair to point 
out that it's not legally binding.

> allow such data in provisionally, i.e. you may contribute the data with
> some sort of flag (no idea about technicalities) that says "if the
> license is ever changed then this data must be removed" or so.

Actually that's not what the current draft says at all.  The draft says the 
data "may" be deleted.  Not "must", not "will be", simply that it "may be". 
This leaves the door open to the fact that the data might not be deleted.

Your above paragrapgh neatly sums up to me why the CT's are incompatible 
with CC-BY, or CC-BY-SA, or indeed many more licences , in that 
compatability of the CT's could only be ensured if:
(a) There was some technical mechanism for fallginf data which needs to be 
removed , and there is no such mechanism; and
(b)  There was a guarntee that usch data "WOULD" be removed, and there is no 
such guarantee.


>> About three or four months ago there was discussion about adding
>> clarification about "free and open license", to add both share alike
>> and attribution clauses. I have two questions - can it still be done,
>> what was working group answer to this, or if not, then why not.
>
> I don't think there's a plan to change the "free and open". You are not
> asking for a clarification, you are asking for additional restrictions
> (as there are many free and open licenses that are not attribution or
> share-alike); such additional restrictions would constitute an undue
> liability for the people who are OSM in the future. (I think it is
> possible that a reference to a widely accepted definition of the terms
> "free" and "open" might be included, as a clarification, but as I said,
> you are not asking for clarification.)
>
> On the whole, if one wants to accept imports that are ODbL comaptible
> but not CT compatible, I think the opt-out version causes less damage
> than trying to toughen up the CT which might cause all kinds of problems
> in the future, but I don't like either.
>
> Imagine our current data came under some sort of CT that said "the
> license may be changed but the new license must be attribution and share
> alike". Now *you* say that ODbL is ok for you, but ODbL does make
> exemptions from share-alike (namely, for non-substantial extracts for
> which attribution and share-alike are dropped, and for produced works,
> for which at leas the share-alike is dropped). I am pretty sure that
> under such hypothetical CTs, a license change to ODbL would not be
> possible.
>
> I think that, even more than "free and open", share-alike is a term that
> is very difficult to define, and if one tries to define it, one will
> already have written half a new license. Licenses interact with their
> legal surroundings; some things we see in ODbL (the non-substantial
> extracts e.g.) are a direct refelction of the database directive. The
> legal surroundings can change, and if we have to change the license in
> the future then a new license might have to reflect the new situation.
> Any sort of cast-in-stone share-alike requirement will be an unnecessary
> burden.
>
> In another post, I have tried to make the point that it would also be
> morally wrong (unfair) against our future colleagues in this project to
> limit their choices.
>
> Since your argument in this posting was not that you personally require
> share-alike but you were concerned only about entering third-party data,
> I would much rather have *less* third-party data and *more* liberty for
> the project in the future, than *more data for the price of reduced
> liberty later. This would be like taking out a mortgage on what OSM is
> in 10 years. We would risk long-term problems for a short-term effect.
>
> Bye
> Frederik
>