Developer discussion for Creative Commons technology and tools

[Mike Linksvayer <ml AT gondwanaland.com>]

Peter has the intent of the vocabulary about right, but of course
unaided machine processing will get lots of cases wrong. So do unaided
humans. A modest but feasible goal is for machines to help humans get
it more right more of the time.

You do need to know the license in order to know how its legal
requirements interact with requested attribution, but
cc:attributionName/URL were intended to interact with license
annotation, with CC license as object use case; otherwise existing DC
terms and others would've been adequate. It may be they're adequate
after all and specifying terms intended to directly facilitate
denoting names and links a licensor wants was foolish. The real world
round trip deployments are lacking to test this, take this as one
will.

It's also worth keeping in mind (as caveat re intent of vocabulary
above) the no restrictions/public domain case. cc:attributionName/URL
might be requests that one cannot be sued for not fulfilling in the
case of a work released with CC0 or annotated with PDM, but a
community may frown on one for not complying with its norms (also see
useGuidelines in the vocabulary). Again it isn't super clear that
specific attributionName/URL terms are the ideal thing here, but I
don't think legal obligation delineates when they are ideal or not
(and if you're tempted to argue that it is the correct delineation,
consider fair use and other exceptions; consider those in your tools,
too).

Mike

On Mon, Jun 24, 2013 at 11:59 AM, Maarten Zeinstra <mz AT kl.nl> wrote:
> Yes you are interpreting CC-REL to narrowly.
>
> CC-REL is used to be able to communicate as basically as possible what a
> person can or cannot do with a license. No RDF-triple, knowledge graph,
> database, rule based systems can be as precise as a legal contract. That is
> because these legal contracts are not meant for machine communication, they
> are meant for natural persons. Only by using very narrow definitions can
> refer to a 'fact' or 'requirement', 'probition', etc. using something like
> RDF. And that definition is possible :) then you need to use <Work>
> <CC:license> <https://creativecommons.org/licenses/by/3.0/> That is the most
> accurate description of Attribution and it needs. Because attribution really
> also depends on the license.
>
> Cheers,
>
> Maarten
>
>
> --
> Kennisland
> | www.kennisland.nl | t +31205756720 | m +31643053919 | @mzeinstra
>
>
>
>
> On Jun 24, 2013, at 18:27 , Peter Liljenberg <peter AT commonsmachinery.se>
> wrote:
>
> No, I don't expect full RDF representation either, just that cc:attribute
> (and the other terms) would be defined as something like "attribute in
> compliance with the CC license legal code", and then guidelines to tool
> developers on what that means (or even just guidelines). This is how I
> started summarising it, but if that is not the intention I appreciate being
> corrected straight away:
>
> The Creative Commons licenses all require attribution, and defines in the
> legal code how to do it. ccREL ties [though it seems not formally] these
> requirements to the metadata on the work, so that if these properties are
> set they must be used in the attribution:
>
> dcterms:license, cc:license or xhtml:license (synonyms in RFD): the URI
> linking to the license terms (e.g.
> http://creativecommons.org/licenses/by/3.0/)
> cc:attributionName: the name of the author and/or designated attribution
> parties
> dc:title: the title of the work
> cc:attributionURL: a URI associated with the work, which should refer to
> copyright or licensing information about the work (otherwise there is no
> obligation to include the URI, and another property should be used)
>
>
> I might be interpreting ccREL too harshly since I'm rather new to this area,
> but this is also because I'm coming from the direction "how can this support
> tooling and automate attribution". This thread indicates that there's a risk
> that a loosely defined cc:attribute (and the other properties too) will
> cause tools to implement incorrect license processing.
>
> /Peter
>
>
> On 24 June 2013 18:12, Nathan Yergler <nathan AT yergler.net> wrote:
>>
>> I don't believe there was any expectation that the RDF representation
>> could fully express the legal code of a license. I think that means
>> Maarten is correct.
>>
>> Of course, there are tools out there that take the attribution
>> requirement and "just happen" to generate attribution text that
>> matches what the CC licenses require. I'd have to think about it more
>> to decide if that's a sane behavior or if they should be checking
>> something else before deciding to do that.
>>
>> NRY
>>
>>
>> On Mon, Jun 24, 2013 at 9:08 AM, Peter Liljenberg
>> <peter AT commonsmachinery.se> wrote:
>> > This reminds me of the question the other week about cc:attributionURL
>> > vs
>> > xmpRights:WebStatement, where the response was that cc:attributionURL
>> > was
>> > related to the legal code of the license even though that wasn't fully
>> > expressed in the ccREL description. That made it map to the semantics of
>> > xmpRights:WebStatement.
>> >
>> > cc:require cc:attribution seems to me to also be related to the legal
>> > code
>> > that specifies exactly what attribution means (e.g. 4b in
>> > http://creativecommons.org/licenses/by/3.0/legalcode). Or is it intended
>> > to
>> > be the more generic term described in the RDF schema?
>> >
>> > /Peter
>> >
>> >
>> >
>> >
>> > On 24 June 2013 17:59, Maarten Zeinstra <mz AT kl.nl> wrote:
>> >>
>> >> Hi Jonas,
>> >>
>> >> To be clear, this is the description of the namespace:
>> >>
>> >> "credit be given to copyright holder and/or author" according to
>> >> view-source:https://creativecommons.org/schema.rdf
>> >>
>> >> So I read this as a binary that when present credit should be given. It
>> >> does not specify a way to do that, and I think it shouldn't as well.
>> >>
>> >> Cheers,
>> >>
>> >> Maarten
>> >>
>> >> --
>> >> Kennisland
>> >> | www.kennisland.nl | t +31205756720 | m +31643053919 | @mzeinstra
>> >>
>> >>
>> >>
>> >>
>> >> On Jun 24, 2013, at 17:36 , Jonas Öberg <jonas AT coyote.org> wrote:
>> >>
>> >> Hi all,
>> >>
>> >> here's a question from IRC which was left hanging. Wondering if anyone
>> >> here has any thoughts about it :)
>> >>
>> >> 09:14 <jonaso> Been looking at
>> >> https://bugs.launchpad.net/inkscape/+bug/372427 which uses ccREL
>> >> cc:permits
>> >> and cc:requires to express licenses which are not CC licenses, ie.,
>> >> FAL.
>> >> They've coded FAL same as CC BY-SA
>> >> 09:14 <jonaso> I wonder if that's the intent: ns#Attribution has a
>> >> specific meaning in the CC vocabulary which is slightly different from
>> >> FAL's
>> >> attribution requirement.
>> >> 09:15 <jonaso> So I wonder if we should think of ccREL ns#Attribution
>> >> as
>> >> "requires some attribution, unspecified exactly how, what or when" or
>> >> if
>> >> ns#Attribution should mean more exactly the terms of the CC licenses.
>> >> 09:16 <jonaso> In the latter case, I guess there should be a separate
>> >> vocabulary to express terms more closely to FAL and other licenses.
>> >>
>> >>
>> >> Sincerely,
>> >> Jonas
>> >> _______________________________________________
>> >> cc-devel mailing list
>> >> cc-devel AT lists.ibiblio.org
>> >> http://lists.ibiblio.org/mailman/listinfo/cc-devel
>> >>
>> >>
>> >>
>> >> _______________________________________________
>> >> cc-devel mailing list
>> >> cc-devel AT lists.ibiblio.org
>> >> http://lists.ibiblio.org/mailman/listinfo/cc-devel
>> >>
>> >
>> >
>> > _______________________________________________
>> > cc-devel mailing list
>> > cc-devel AT lists.ibiblio.org
>> > http://lists.ibiblio.org/mailman/listinfo/cc-devel
>> >
>
>
> _______________________________________________
> cc-devel mailing list
> cc-devel AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/cc-devel
>
>
>
> _______________________________________________
> cc-devel mailing list
> cc-devel AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/cc-devel
>