Thanks, Robla and Nemo!

We will bring up these points when we meet with the wikidata team in a week to discuss the Structured Data project (1).

My initial impression is that this would be a significant scope increase for an already large project. So for practical reasons, we may not be able to take it on until after we have implemented structured data for multimedia files.

That said, it would be important to consider this request before we start development on structured data for multimedia files, so we can investigate solutions that could make this second phase possible. So we will add this discussion to our meeting agenda.

We will update the relevant pages after we’ve had a chance to discuss this as a team.

To be continued,


Fabrice


P.S.: I have also Cc:d Luis and Stephen from our legal team, so they can remain part of that discussion as well.


(1) https://commons.wikimedia.org/wiki/Commons:Structured_data



On Sep 26, 2014, at 2:37 PM, Federico Leva (Nemo) <nemowiki@gmail.com> wrote:

Rob Lanphier, 26/09/2014 22:59:
Thanks for the analysis, Gergo!  I was going to split Luis' proposal
into a separate wiki page, but I see Nemo has linked to this page as the
"Canonical page on the topic":
https://www.mediawiki.org/wiki/Files_and_licenses_concept

Without a deep reading that I'm admittedly just not going to have time
for, it's hard to tell how related the page that Nemo linked to is to
the concepts that Luis is trying to capture.

The gist of that idea is: associate actual copyright metadata to content; use ContentHandler for certain blobs of information. Krinkle and others were the main authors of that page and the idea was never worked on, but it can be extended in many ways... except that it's a bit pointless to expand it further when even a smaller scope is hard to work on.

Other than files, the two classic pain points about copyright metadata are
a) display of page authors https://bugzilla.wikimedia.org/show_bug.cgi?id=2994#c14
b) metadata about works (e.g. books) stored across pages https://bugzilla.wikimedia.org/show_bug.cgi?id=15071

Could someone (Nemo?
Luis?) merge Luis's requirements into the "canonical page" to Luis'
satisfaction, so I can delete most of the information from our backlog?
I'll keep the item on the MW Core backlog, since I don't know where else
to put it, but it's probably going to be relatively low priority for
that team.

Multimedia team and Wikidata team, could you make sure you're
considering the requirements that Luis brought up as you build your
solution?  Even if you decide to punt on some of the things that aren't
strictly necessary for files, it's still good to make sure you don't
paint us in a corner when if/when we do try to do something more
sophisticated for articles.

A solution based on an external wiki (Wikidata) as for files... may work for (b) but won't for (a). That said, the original idea for files could be reused for both (a) and (b).


One thing I'll note, though, before we get too complacent in thinking
that files are somehow simpler than articles, we should consider these
relatively common scenarios:
*  Group photo with potentially different per-person personality rights
*  PDF of a slide deck with many images
*  PDF of a Wikipedia article  :-)

The last point being bug 2994 (and friends).

Nemo

_______________________________________________
Multimedia mailing list
Multimedia@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/multimedia

_______________________________

On Sep 26, 2014, at 1:59 PM, Rob Lanphier <robla@wikimedia.org> wrote:

(+cc Nemo and Wikidata-tech)

On Fri, Sep 26, 2014 at 5:33 AM, Gergo Tisza <gtisza@wikimedia.org> wrote:
On Fri, Sep 26, 2014 at 2:49 AM, Rob Lanphier <robla@wikimedia.org> wrote:
There's an item that's Luis Villa added to the MW Core backlog that I'd like to move to the Multimedia backlog:

I'm assuming everything that he describes fits nicely into what is planned for Structured Data.  Assuming that's true, should I just copy/paste into a new card in Mingle, or a new page on mw.org or what?

This seems to be about article text, or mainly about article text (articles imported from other wikis and so on).

The plan for the structured data project is to create Wikidata properties for legalese, install Wikibase on Commons (and possibly other wikis which have local images), make that Wikibase use Wikidata properties (and sometimes Wikidata items as values), create a new entity type called mediainfo (which is like a Wikibase item, but associated with a file), and add legal information to the mediainfo entries.

Part of that (the Wikidata properties) could be reused for articles and other non-file content - the source, license etc. properties are generic enough. However, if we want to use this structure to attribute files, we would either have to make mediainfo into some more generic thing that can be attached to any wiki page, or abuse the langlink/badge feature to serve a similar purpose. That is a major course correction; if we want to do something like that, that should be discussed (with the involvement of the Wikidata team) as soon as possible.

Thanks for the analysis, Gergo!  I was going to split Luis' proposal into a separate wiki page, but I see Nemo has linked to this page as the "Canonical page on the topic":

Without a deep reading that I'm admittedly just not going to have time for, it's hard to tell how related the page that Nemo linked to is to the concepts that Luis is trying to capture.  Could someone (Nemo? Luis?) merge Luis's requirements into the "canonical page" to Luis' satisfaction, so I can delete most of the information from our backlog?  I'll keep the item on the MW Core backlog, since I don't know where else to put it, but it's probably going to be relatively low priority for that team.

Multimedia team and Wikidata team, could you make sure you're considering the requirements that Luis brought up as you build your solution?  Even if you decide to punt on some of the things that aren't strictly necessary for files, it's still good to make sure you don't paint us in a corner when if/when we do try to do something more sophisticated for articles.

One thing I'll note, though, before we get too complacent in thinking that files are somehow simpler than articles, we should consider these relatively common scenarios:
*  Group photo with potentially different per-person personality rights
*  PDF of a slide deck with many images
*  PDF of a Wikipedia article  :-)

Rob