[od-discuss] Please review: 'Datenlizenz Deutschland' for OD compliance.

[Mike Linksvayer]

Thanks for driving this forward Daniel!

I made some substantial edits, reflecting my opinion that the problem
is clarity, but not that any of the likely readings make it non-Open.

I also elided some points that might be re best practices, but IMO it
isn't entirely clear what the best practice in each case is, and they
don't have any bearing on conformance. Coped below with some of
Daniel's text and my comments. I think if we keep these in any form
they should be in a postscript/addendum and clearly demarcated as not
bearing on conformance:

1. it doesn't explicitly make sure how to reference the actual data
holder. (Meaning how to "mention the source"? I'm not sure this is a
problem. Overspecification can be more of a problem than
underpecification of attribution.)
2. it doesn't explicitly make sure how to reference the license text
itself. (This is a minor point, but could be more explicit, maybe
provide canonical URI for refernecing license? This has nothing to do
with being conformant or not.)
3. it doesn't explicitly make sure that it will be compatible with all
further versions of the license. (Why should it? It has no copyleft
aspect.) << this means that content that is licensed withv1.0 can
easily be upgrated to v1.1 ... Yes I know there's a case for such, but
it is an edge case unsupported by all(?) permissive licenses to date;
I don't know why they should.



Tell me I'm wrong!
Mike

On Mon, Mar 18, 2013 at 8:58 AM, Daniel Dietrich
<daniel.dietrich at okfn.org> wrote:
> Dear all,
>
> thanks to all for your comments.
>
> I would like to consolidate this into a short replay to the people in charge at the Ministry of the Interior as they requested the Advisory council to check the license for compliance with the OD.
>
> I have booted this pad to draft this answer:
>
> http://opendefinition.okfnpad.org/DE-OGD-license
>
> Please help me to get this done, so we can send it out asap. Thanks!
>
> Kind regards
> Daniel
>
>
>
>
> On 3 Mar 2013, at 00:21, Mike Linksvayer <ml at gondwanaland.com> wrote:
>
>> On Thu, Feb 28, 2013 at 9:07 PM, Andrew Stott
>> <andrew.stott at dirdigeng.com> wrote:
>>> It’s unclear to me what the precise meaning of the requirement is.
>>>
>>> (1) is it saying (a) that all the actual changes must be documented (Herb’s
>>> interpretation); or is it saying (b) that the attribution statement must be
>>> different if the data is changed (for instance from “reproduces data from
>>> Ministry” to “modified from data from Ministry”)?  The original German
>>> refers to a “Veränderungshinweis” which Google translates as a “Change
>>> Notice”, which could mean either a description of the changes or simply a
>>> notice that changes had been made.
>>>
>>> I recall seeing an Australian example [1] of CC-BY which required a
>>> different attribution statement if that data had been changed.
>>>
>>> (2) is it saying (a) that the licensor can later at its discretion instruct
>>> the licensee to remove the attribution (Herb’s interpretation); or is it
>>> saying (b) that the licensor can specify *in advance* that an attribution
>>> should *never* be given if the data is changed?
>>>
>>> I agree with Herb that if the meaning is 1(a) and 2(a) then it is
>>> non-conformant – and 1(a) would be burdensome and 2(a) would leave
>>> indefinite uncertainty.
>>>
>>> However requiring simply a different fixed attribution text (as in
>>> interpretation 1(b)), or no attribution at all (2(b)), if the data has been
>>> changed would seem to me to be a conformant way of the licensor distancing
>>> himself from modified data where there was reputational or moral liability
>>> involved.
>>
>> I'm not sure what the clause means either, but my best guess is a weak
>> version of 1(a) together with 2(b) (referring to Andrew's descriptions
>> above, not license sections). By weak, I mean that it doesn't require
>> a machine-readable version of changes. It would be great if the
>> language were more explicit, or if it already is, a German speaker
>> could explain.
>>
>> But I'm not sure any of above readings would make the proposed
>> attribution license non-Open. 2(a) is problematic, but then all
>> previously-deemed-Open CC BY and BY-SA licenses have that. As they do
>> require a weak version of 1(a). And previously-deemed-Open licenses
>> includes a strong version of 1(a), in particular ODbL does, as an
>> alternative to full source distribution (which may or may not be less
>> onerous depending on circumstances; OTOH maybe as an alternative it is
>> OK, but without full source option it is not).
>>
>> Though I'm not certain above readings would make proposed license
>> non-Open, I'd recommend both making conditions more clear and ensuring
>> they are minimal -- this presumably aims to be a permissive
>> attribution-only license; its effectiveness as such will be greatly
>> limited by difficult to understand and possibly understood as onerous
>> conditions.
>>
>> Mike
>