[OSM-legal-talk] Share-Data Licenses and OSM
80n80n at gmail.com
Wed Feb 28 15:12:28 GMT 2007
On 2/28/07, Frederik Ramm <frederik at remote.org> wrote:
> advocates of CC-BY-SA say that SA is a way to ensure that
> something is given back to OSM from people building on top of OSM.
> Others have pointed out that the "SA" only affects the finished
> product, not the intermediate geodata, and thus might not be worth a
> lot to OSM (example: guy makes copy of OSM London map data, spends
> two years refining it, publishes printed atlas - only the printed
> atlas must be shared, not his data). There have been calls to extend
> or modify our license to include a "share-data" aspect, i.e. asking
> for any "geodata" created on top of OSM to be published/copylefted.
> (For the record, I am still in favour a full PD approach, but that
> doesn't prevent me from thinking about other models and how they
> would work.)
> Let us assume for a moment that we had such a share-data license in
> place. How would that work? Assume someone makes a copy of OSM data
> and makes changes to it - correcting errors, adding his own tags or
> landmarks, whatever. He is, of course, not bound by any tagging rules
> that we employ for ourselves; what he creates will, in any but the
> most trivial cases, be incomaptible with OSM. In addition, OSM data
> also changes. Half a year later, the guy publishes his work, and duly
> makes his source geodata publicly availalbe.
> What use will there be for OSM? If you gave me last month's planet
> file augmented by 100 segments and asked me to somehow "recapture"
> those into today's OSM, that would already be extremely difficult
> (how am I to tell whether a segment found in your planet file and not
> in our data was added by you and not just deleted by an OSM user?).
> Give me a planet file where you have made a lot of corrections, it
> becomes more difficult. Add your own tagging, and I'll have even more
> problems. Or publish the data in a format with higher complexity
> (adding information about lanes, turn restrictions, whatever), and
> I'll be completely screwed - your data simply cannot be put in OSM.
It may be that this person's work would not be of use to OSM, but I
think that is only half of the reason for the SA clause. Because OSM
has the *right* to reincorporate the additional data does not
necessarily mean that all additional data will be of sufficient
quality or kind to reincorporate it.
But what it does mean is that if, say, the Ordnance Survey took the
OSM dataset and combined it with some other data to make a really nice
map of Yorkshire then we would have the *right* to have that data
back. This prevents anyone from using the OSM data as a free launch
pad for their own competing product unless it is made available on the
same license terms, in which case OSM gets it back.
One of the very interesting things about the SA clause is that as much
as it encourages giving back it also prevents fragmentation and
This article, referring to GPL vs BSD alludes to this kind of effect:
> Not that such a publication would not be interesting to some "as is",
> if only to re-create modified version of whatever work was produced
> by the person. But I cannot for the life of me think of how we could
> ever capture the essence of such changes back into OSM, especially if
> whoever created the data does not cooperate.
> My conclusion is that if we were to implement a share-data license
> with the express aim of OSM improvement by "getting something back",
> then "getting something back" must be very clearly defined, because
> just by publishing their derivates created from OSM data, nobody
> helps OSM.
> Frederik Ramm ## eMail frederik at remote.org ## N49°00.09' E008°23.33'
> legal-talk mailing list
> legal-talk at openstreetmap.org
More information about the legal-talk