Showing posts with label curation. Show all posts
Showing posts with label curation. Show all posts

Saturday, July 14, 2018

#AfricaGap - Where Wikipedias collide

The German and the English Wikipedia collide on the "administrative territorial entities" of the Gambia. I was told to remove entries that I made to Wikidata because they were "Falschinformationen". The German article is much better written but the English article indicates that the German information is likely to be outdated.

A discrepancy like this is obviously best solved to insist on "your" solution. The point that I have been making quite often is that such differences are commonplace and require proper sourcing. The obvious source will not be found on a university website, it will be found in governmental information of the Gambia.

Making information about Africa available in Wikidata makes the errors, the inconsistencies and the lack of data in the Wikipedias more visible. This is not solved by considering your "own" data to be best, it is by proving that information is up to date. According to the English Wikipedia, the Upper River Division is no longer; it is largely replaced by the Basse Local Government Area.

My question: what does it take for the Wikipedias to take their inconsistencies serious?
Thanks,
      GerardM

Sunday, October 22, 2017

#Katherine - When the Cebuano issue is no longer about #Wikipedia

Dear Katherine, I loved your presentation at the Berkman Klein Center for Internet and Society. It has much to think about and <grin> it is great that you answer the question you want to answer </grin>.

You address questions like "will we let external organisations use our data for their own purposes". My suggestion to you, us all, is why not use our own data for our own purposes.

The Cebuano Wikipedia is seen as problematic on many levels. It is one of the biggest Wikipedias in number of articles and one of the smallest in the size of its community. Like any Wikipedia, its articles are harvested for use in Wikidata and that brings us to several problems but more importantly in the light of your presentation, opportunities.

Problem: the data used the Cebuano articles are based is problematic
Opportunity: import the data in Wikidata first and first do some curation there.

Problem: the data is licensed under a CC-by-sa license and Wikidata is CC-0
Opportunity: collaborate with the copyright holder and ask their permission to include the data in Wikidata

Problem: when text is generated by a bot, the text when saved in an article is fixed
Opportunity: do not save it as an article but generate the text and maybe cache the text

Problem: other organisations use our data to generate information
Opportunity: we generate information in all the 300 languages where Wikipedia does not have an article

Problem: we have information that has no article in any language
Opportunity: we generate the text and maybe cache the text

Problem: Wikimedia officials indicated that issues like the Cebuano Wikipedia are not relevant
No opportunity; opportunities for all our projects are missed

Katherine, we already generate texts using bots, we already cache our data, we do it for English, we do it for Swedish, Cebuano. Why leave it for the companies of our world to generate text where there is already so much? We can do better, do the same and do it for all our languages as well.
Thanks,
      GerardM