Most referenced domains on the Swedish Wikipedia (2015-04-15) (T96927)

ActivePublic

Actions

@gpaumier : I belive this is not correct, it is at least two (perhaps even three) orders of magnitude too low. How did you get these numbers?

@Ainali: The task is referenced above (T96927) and there is also more information at m:Research:Citoid support for Wikimedia references.

The numbers did seem low to me, but I couldn't find a reason why at first glance. Feel free to correct the list (or create another paste).

Well, I try to read the links you provide, but I fail to understand what you try to count. Is it the number of external links encapsulated by a <ref> tag?
If so, this search will give you plenty of hits on that criteria for a domain that did not make the list: https://sv.wikipedia.org/w/index.php?target=*.catalogueoflife.org%2F&title=Special%3AL%C3%A4nks%C3%B6kning

I think there may be a problem with this specific list, as at some point it seems to be following an alphabetical order rather than a relevance one?

The only reason I suspect that we could be a whole order of magnitude off is that we aren't matching a ton of legitimate references. Is it possible that most references do not have URLs in them or that Swedish Wikipedia doesn't use <ref> tags as much as other wikis?

If neither of those makes sense, I'd like to manually process a few revisions. I added a feature request (https://github.com/halfak/mwrefs/issues/3) to allow us to run tests against specific pages using the API. Could you give me a sense of the urgency so I can prioritize the work?

This is a typical bot created butterfly: https://sv.wikipedia.org/wiki/Phrixocomes_gephyrea
It has the Catalouge of life URL in a <ref> tag, which also the common way to add references on Swedish Wikipedia.

Aaron, I have requested a couple of similar lists for Dutch and German (T107788, T108689), and the Wikipedia Library has expressed interest in these reports (see T96927), so maybe we want to figure out if there's anything wrong with the script or something before someone generates other lists which may turn out to be "inaccurate". Thank you!

@Elitre, the first step is to *show* that there is an issue and where it appears. Since @gpaumier owns the pipeline, I'll need him to look into where it seems that we are getting low counts. If it can be shown to be a breakdown in mwrefs, I'll be happy to take a look at it.

@Ainali, see https://sv.wikipedia.org/wiki/Anv%C3%A4ndare:Edgars2007/Most_frequent_domains . What do you think about this one?

@Elitre that looks much more like what I would have expected.