It looks like a solution to bug 4547 is on the horizon.
https://bugzilla.wikimedia.org/show_bug.cgi?id=4547
See also [Wikitech-l] Reasonably efficient interwiki transclusion http://www.gossamer-threads.com/lists/wiki/wikitech/197322
This will be very useful for templates which Commons has developed, especially language related templates, however I am concerned that people are also planning on using Commons as a repo for Wikipedia infoboxes, and including the *data* on Commons rather than just the template code. e.g.
http://www.mediawiki.org/wiki/User:Peter17/GSoc_2010#Interest
This centralisation of data makes sense on many levels, however using Commons as the host of this data will result in many edit wars moving to the Commons project, involving people from many languages. Even the infobox structure can be the cause of edit wars.
I think it is undesirable to have these Wikipedia problems added to Commons existing problems. ;-)
Tying Wikipedia and Commons closer together is also problematic when we consider the differing audience and scope of each project, especially in light of the recent media problems. If the core templates and data used by Wikipedia are hosted/modified on Commons, it will be more difficult to justify why Commons accepts content which isn't appropriate on Wikipedia.
A centralised data wiki has been proposed previously, many times:
http://meta.wikimedia.org/wiki/Wikidata/historical http://meta.wikimedia.org/wiki/Wikidata http://meta.wikimedia.org/wiki/Wikidata_%282%29 http://meta.wikimedia.org/wiki/WikiDatabank
Non-WMF projects, such as freebase, dbpedia, etc., have been exploring this space.
Isn't it time that we started a new project!? ;-)
A wikidata project could use semantic mediawiki from the outset, and be seeded with data from dbpedia.
A lot of existing & proposed projects would benefit from a centralised wikidata project. e.g. a genealogy wiki could use the relationships stored on the wikidata project. wikisource and commons could use the central data wiki for their Author and Creator details.
-- John Vandenberg
A wikidata project could use semantic mediawiki from the outset, and be seeded with data from dbpedia.
A lot of existing & proposed projects would benefit from a centralised wikidata project. e.g. a genealogy wiki could use the relationships stored on the wikidata project. wikisource and commons could use the central data wiki for their Author and Creator details.
+1
Could this be part of dbpedia?
SJ
(also including foundation-l as this isn't really a commons-specific discussion)
On 22 Nov 2010, at 21:04, Samuel Klein wrote:
A wikidata project could use semantic mediawiki from the outset, and be seeded with data from dbpedia.
A lot of existing & proposed projects would benefit from a centralised wikidata project. e.g. a genealogy wiki could use the relationships stored on the wikidata project. wikisource and commons could use the central data wiki for their Author and Creator details.
+1
Could this be part of dbpedia?
dbpedia is about collating the information available on Wikipedia and providing that as a database for others to use. This is about having a central information store that can be edited to add information. Whilst dbpedia could seed wikidata, they're very different projects in the way they would operate.
In my opinion, the Wikimedia Foundation should very seriously look into starting something like wikidata. I don't suppose there's a facilitator that could be hired that knows about Wikimedia sufficiently to facilitate an on-wiki discussion and formation of a comprehensive proposal to start this project, including bringing together the various people interested in this project?
Mike Peel
On 11/22/2010 10:24 PM, Michael Peel wrote:
In my opinion, the Wikimedia Foundation should very seriously look into starting something like wikidata.
One major problem is that different people have very differing understanding of what "wikidata" should mean. It is an abstract good, similar to "world peace" or "democracy" -- or Wikimedia's "usability" project, that introduced slow and broken Javascript instead of actually increasing usability. With this background I would advise against starting a "wikidata" project.
What should be started is something smaller and more focused, that solves some actual problem. This is like asking for "freedom of the press" or "women's suffrage" rather than abstract "democracy".
So, which concrete, smaller ambitions could you list?
From my very personal and (GLAMmy) point of you, it should be, at least,
a OAI-MPH (OAI-ORE?) complaint repository, who would harvest/contain metadata from Commons and Wikisource. It should import/export data in variuos forms and within various metadata schemes (e.g. Dublin Core).
It could be designed for llong-term preservation, adding archival metadata (and metadata schemes).
For the part of Wikidata as a repository of Open Data (Linked Open Data?) from Wikipedia, I give you the word.
Aubrey
2010/11/24 Lars Aronsson lars@aronsson.se
On 11/22/2010 10:24 PM, Michael Peel wrote:
In my opinion, the Wikimedia Foundation should very seriously look into
starting something like wikidata.
One major problem is that different people have very differing understanding of what "wikidata" should mean. It is an abstract good, similar to "world peace" or "democracy" -- or Wikimedia's "usability" project, that introduced slow and broken Javascript instead of actually increasing usability. With this background I would advise against starting a "wikidata" project.
What should be started is something smaller and more focused, that solves some actual problem. This is like asking for "freedom of the press" or "women's suffrage" rather than abstract "democracy".
So, which concrete, smaller ambitions could you list?
-- Lars Aronsson (lars@aronsson.se) Aronsson Datateknik - http://aronsson.se
Commons-l mailing list Commons-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/commons-l
On 11/23/2010 10:01 PM, Lars Aronsson wrote:
On 11/22/2010 10:24 PM, Michael Peel wrote:
In my opinion, the Wikimedia Foundation should very seriously look into starting something like wikidata.
One major problem is that different people have very differing understanding of what "wikidata" should mean. It is an abstract good, similar to "world peace" or "democracy" -- or Wikimedia's "usability" project, that introduced slow and broken Javascript instead of actually increasing usability. With this background I would advise against starting a "wikidata" project.
What should be started is something smaller and more focused, that solves some actual problem. This is like asking for "freedom of the press" or "women's suffrage" rather than abstract "democracy".
So, which concrete, smaller ambitions could you list?
Sure, but the problem small steps only address small problems and do not build a foundation to fix larger issues. A series of template hacks, bots and external indexers could address most small ambitions, but a unified underlining structure would bring many greater 'ambitions' within reach. For example if we just wrote a one off system to exclusively handle interlanguage links, it would not be nearly as useful as a system that tied these inerlanguage links together as multilingual labels for relational data. ( Apples are fruits, even when they are 'Manzanas' and 'Fruta', something that would be wholly lost in the 'one off' solution.
To borrow from your analogy you need a "constitution" before "freedom of the press" or "womens suffrage" makes any sense, else every movment is extrodinary laborious, can't build on any other effort and is exclusive to a single problem for a single set of people.
--michael