[cc-community] Geograph British Isles publishing entire cc-licensed archive via bittorrent

Mike Linksvayer ml at creativecommons.org
Sat Jul 5 01:39:03 EDT 2008


On Fri, Jul 4, 2008 at 1:46 AM, Paul Dixon <paul at elphin.com> wrote:
> Geograph British Isles (http://www.geograph.org.uk) collects over a
> thousand geolocated, cc-licensed pictures of the British Isles every day
> - we've amassed over 860,000 so far.

This is fantastic!

Just a few comments on the metadata, torrents, and site.

Least important first.  Here's some metadata from the sample torrent:

    <rdf:Description rdf:about="00/00/000006.jpg">
    <dc:identifier>http://www.geograph.org.uk/photo/6</dc:identifier>
    <dc:title>Arches near Gloucester Cathedral</dc:title>
    <dc:subject>Arch</dc:subject>
    <dc:creator>Helena Downton</dc:creator>
    <dc:dateSubmitted>2005-03-06T10:13:10Z</dc:dateSubmitted>
    <georss:point relationshiptag="is-picture-of">51.8670807236559
-2.24547177500684</georss:point>
    <dc:licence>http://creativecommons.org/licenses/by-sa/2.0/</dc:licence>
    <dc:format>image/jpeg</dc:format>
    <dc:type>http://purl.org/dc/dcmitype/StillImage</dc:type>
    <dc:publisher>www.geograph.org.uk</dc:publisher>
    </rdf:Description>

There are a bunch of URI values there, but to a RDF processor, they
are text literals.  Shouldn't the values be in rdf:resource
attributes, like

    <rdf:Description rdf:about="00/00/000006.jpg">
    <dc:identifier rdf:resource="http://www.geograph.org.uk/photo/6"/>
    <dc:title>Arches near Gloucester Cathedral</dc:title>
    <dc:subject>Arch</dc:subject>
    <dc:creator>Helena Downton</dc:creator>
    <dc:dateSubmitted>2005-03-06T10:13:10Z</dc:dateSubmitted>
    <georss:point relationshiptag="is-picture-of">51.8670807236559
-2.24547177500684</georss:point>
    <dc:license rdf:resource="http://creativecommons.org/licenses/by-sa/2.0/"/>
    <dc:format>image/jpeg</dc:format>
    <dc:type rdf:resource="http://purl.org/dc/dcmitype/StillImage"/>
    <dc:publisher rdf:resouce="http://www.geograph.org.uk"/>
    </rdf:Description>

?

And I hate to say it, but there is no such thing as dc:licence, only
dc:license. In fact, there is no dc:license either given
xmlns:dc="http://purl.org/dc/elements/1.1/".  license (and
dateSubmitted) are only available in an extended DC specification, see
http://dublincore.org/documents/dcmi-terms/ -- fortunately entirely
backwards compatible, so you'd just need to change the xmlns:dc value
to http://purl.org/dc/terms/



Second, http://torrents.geograph.org.uk/ says

    All of the data is licenced under a Creative Commons licence,
    and the RDF file references the licence terms for each item in the
    volume. Please take care to respect these licences when re-using
    this data!

I suppose technically "data" can mean everything in the torrent --
images and metadata, but I don't know if that will be obvious to many
people.  And since everything is under Attribution-ShareAlike, it
would be good to specify that so people don't have to download the
torrent just to find out. :)  That could be done with a link if you
don't want to spell it out in text.  So I would suggest something like
this

    Everything in the torrents -- images and
    metadata -- is licenced under the <a rel="license"
    href="http://creativecommons.org/licenses/by-sa/2.0/">Creative Commons
    Attribution-ShareAlike licence</a>, and the RDF file references
    the licence terms for each item in the volume. Please take care to
    respect these licences when re-using this data.


Finally, it would be great to call out the specific license used
elsewhere on the site -- the actual image pages are perfect, linking
to the exact license used -- I'm thinking of especially
http://www.geograph.org.uk/faq.php#opensource and
http://www.geograph.org.uk/help/freedom both of which mention a CC
licence, even one that allows commercial use, but don't say which one.


Of course if not everything is under the same licence then two and
three above would have to be modified.

Note that I'm not complaining, these are really minor suggestions, to
what as I lead with above is a fantastic project!

Mike


More information about the cc-community mailing list