[OSM-talk] Opinion poll about the new licence Odbl 1.0
frederik at remote.org
Sun Dec 6 21:39:13 GMT 2009
Ævar Arnfjörð Bjarmason wrote:
> On Sun, Dec 6, 2009 at 20:36, Liz <edodd at billiau.net> wrote:
>> For Australians it means the loss of the coastline, most of which has been re-
>> edited from government data, and major rivers like the Murray
> If someone presents me with a boolean "Do you allow relicensing under
> the ODbL" I'll have to say no because some of my edits are derived
> from CC-BY-SA data I don't have permission to license (and I probably
> can't even recall what all of it is).
> Which'll mean nuking >50% of all the data in Iceland most of which
> I've touched at some point.
First, I would appreciate if people could stop talking about "nuking" data.
The absolute worst case, where data cannot be re-licensed into ODbL
because the original contributor is dead, does not agree, cannot be
reached, or cannot be bothered to read our proposal, is this:
The non-relicensed data will sit in some kind of separate, possibly
read-only server, from where it can be accessed, just like now, under
the terms of CC-BY-SA. This server may or may not be made available by
OSMF but it will certainly exist, and OSMF has already said that a full
history dump will be provided.
We will, in all likelihood, be rendering tiles that display the old data
alongside the new data in a fashion largely indiscernible from today's
maps. These map tiles will have to be CC-BY-SA licensed (because part of
them comes from CC-BY-SA sources) but that's fine with us. (OSMF has not
made a statement, and probably neither a decision, about whether or not
the osm.org tileserver will serve such mixed renderings but if that
server doesn't then you can be sure others will fill the need.)
We might even - and again, this is something outside of OSMF's control
and can be set up by any interested group in the project - allow limited
write access to the old CC-BY-SA database, so that when things are
eventually relicensed or resurveyed, they can be removed from the old
data set to avoid rendering conflicts.
So for map rendering, the damage will be, I shall say, minimal. More
effort for rendering, yes, but the same good maps that we already have.
It will be more difficult for routing engines or other users of our data
because combining CC-BY-SA and ODbL data in a database is not possible
except in fringe situations where you can get away with having a
Also, of course, editing will be more difficult because you have the
legacy data. But even here it is thinkable to have editors that will
download old and new data, and maybe display the old data in a "greyed
out" version or so, indicating that editing is only possible on the new
data. (There's neither technical nor legal reason to disallow editing on
the old data, but we do want to have an incentive for people to
ultimately make the switch I think. Also we have to be careful not to
copy data from one dataset to another.)
But this is the worst case. I firmly believe that it will be possible to
come to terms with many contributors, even if they disagree with ODbL at
the moment, or if they are government bodies which act at turtle speed.
It will take some effort and may not always work, but I see no reason to
be so pessimistic about this. (It will be necessary for OSMF to rein in
those in it's ranks who think that this can be achieved by insulting
anyone who is against ODbL, but I trust this will automatically come as
the organisation matures.)
Also, there will surely be a fine-grained approach to edits. Just
because you have touched something in Iceland and cannot make the switch
to ODbL, one can still retrieve the version from before you touched it,
and use that. Better than nothing.
(In cases like yours, I think one should really make an effort to
determine which of your edits are "tainted" by external CC-BY-SA
sources. I think it would be ok to get this 90% right, it doesn't have
to be absolutely correct - if a few CC-BY-SA items slip through, or if a
few non-CC-BY-SA items get dropped, the damage isn't that big.)
Frederik Ramm ## eMail frederik at remote.org ## N49°00'09" E008°23'33"
More information about the talk