[od-discuss] Great post on difficulty of actually using "open" data

Tue Oct 29 20:26:13 UTC 2013

On Sat, Oct 26, 2013 at 7:15 PM, Luis Villa <luis at lu.is> wrote:
> On Wed, Oct 23, 2013 at 12:49 PM, Mike Linksvayer <ml at gondwanaland.com>
> wrote:
>> > That's not a problem with copyleft, although writing and using too many
>> > incompatible copyleft licenses is a problem.
>>
>> I agree, don't think the example demonstrates a problem with copyleft.
>> It is barely pertinent. We could have same old debate about which
>> effect would be greater: more use of copyleft would cause more of
>> those non-open datasets to be opened, or more use of copyleft would
>> make open datasets harder to use, less valuable. But it would be a
>> hypothetical argument based on the example provided.
>
> I raised copyleft because OSM, and in the near future many CC 4
> gov't-created datasets, are some of the largest "open" datasets around. And
> I'm quite comfortable saying that ODbL, and CC4, are going to cause a lot of
> headaches when people try to use them in this sort of scenario. So no,
> copyleft isn't mentioned in this particular blog post, but it is a very
> real/non-hypothetical concern.

Real concern, hypothetical net effect.

By the way, I noticed another case which you may already be familiar
with, but probably others aren't: Mozilla's new cell tower/access
point location service and existing databases, links in
https://news.ycombinator.com/item?id=6630010

https://wiki.mozilla.org/Services/Location/Bootstrap is quite funny:
"Some data are shared between projects, for example OpenBmap [ODbL 1.0
& CC-BY-SA 3.0] imports data from OpenCellID [CC-BY-SA 3.0] on a
regular basis and Enaikoon [CC-BY-SA 3.0] started based on an older
snapshot of opencellid.org data." The table also notes OpenWLANMap,
which is GFDL, and several others with no license or very limited
non-commercial only licenses.

The funny bits are a GFDL dataset and a ODbL & BY-SA 3.0 dataset that
imports data from a BY-SA 3.0 one. Of course the latter may be OK;
perhaps the "import" and subsequent uses don't mix the data at all, or
perhaps copyright is definitely not pertinent, in which case BY-SA 3.0
is not pertinent, and those databases may be treated as public domain.
But I wouldn't be on either of these no-problem cases (I've not looked
further at all).

Skimming issue comments and a couple list threads, I guess there are
plenty of privacy and technical issues to sort out before Mozilla
decides to release any data, and whether any of the existing databases
would be useful, and further what terms (I see CC0 and ODbL+DbCL
mentioned) to release any eventual data under.

Mike