Authorship

The German Wikipedia (but apparently not the English one) has a feature to assess authorship of individual articles. Clicking on "Autoren" at the foot of any articles directs you to a page like this one: https://xtools.wmflabs.org/articleinfo-authorship/de.wikipedia.org/Collectio_Dacheriana?uselang=de where every author is attributed a certain percentage of the current text. Would that be attractive for us, too? --Christof Rolker (talk) 10:50, 28 September 2022 (CEST)

The tool as it is works only with Wikipedias, but the code is available on GitHub and may be used for our Wiki too. However, the most important authorship is the 'main author' as found in the new infoboxes. A tool like the XTools authorship tool may help to establish this, but in the end we do not need it. It is not always important who contributed most characters, but who is responsible for important changes/contant, so the 'main authr' will have to be determined by different means anyhow. Implementing the XTools authorship would be nice at some time, but certainly not important for the next years. --Christof Rolker (talk) 17:49, 4 July 2023 (CEST)

We now have a Wiki user called "Linda Fowler-Magerl" and we have attributed the initial uploads of the descriptions (which were taken from Linda's book) to this user. Thus, if we ever get to install one of the authorship tools, this will give us a more realistic picture of each author's contribution. Originally the upload was done by a program in my name, which of course distorted the picture considerably. --Clemens Radl (talk) 10:27, 3 May 2024 (CEST)

Note: The initial upload was done in several passes by my program. Usually those changes were marked with the same comment ("Initial upload from book."). I used this comment to identify Linda's contributions. In some cases this may have missed some contributions as my script sometimes operated without leaving a comment (e.g., see The first canon law collections in the Latin West). I think, this problem is not a big one and can be tackled with some database wrangling, but we have to keep it in mind. --Clemens Radl (talk) 10:48, 3 May 2024 (CEST)

I did a bit of digging: As was already stated, the XTool (or, because it's a suite, more precisely the tool "authorship" of the suite) that is used on Wikipedia has not been supported for non-Wikimedia wikis since 2021 (source, the last supported version 3.11.2 see here). Unfortunately, it's not as simple as installing an older version of the suite and disabling all other tools that we don't need because the authorship-tool relies on third-parties which very likely won't be compatible with the older version (see also one of the dev's comment on this here). I don't think the benefit of trying and adjusting everything that it'd work for us would be worth it, not least because we'd also have to maintain the code and update it regularly.

I have also not managed to find another tool with similar functionality, and seeing as XTools dropped the support for non-Wikimedia wikis because there was not enough demand for the whole suite on other wikis, I think it unlikely that I simply missed one.

One option would be to implement it ourselves in a way somewhat similar to the infoboxes. It's relatively easy to get all revisions for all pages with all of their authors by using the API (or a bot). That data in turn could be used to calculate which user contributed how many bytes for each page. This could be regularly inserted into a template (it's probably possible to only update if changes were made). This kind of code can be written relatively easily, probably max. 10 hours (from my side, I don't know long it would take to set the regular updates up).

This obviously has a few drawbacks though: The date would need to be collected and updated regularly (same as with the infoboxes) which could ofc mean that the template is out of date. Secondly, the data would be less accurate than the one displayed by the authorship-tool (not every byte needs to be a character, e.g., it could also be an image, and some characters are more than one byte long, etc.).--SStark (talk) 20:58, 13 May 2024 (CEST)

Images are not really a problem, and number of bytes is precise enough for our purposes (on which see below). --Christof Rolker (talk) 12:20, 14 May 2024 (CEST)

In the end it probably depends on what our goal is. Do we want articles to be citable in individual cases? If that's the case I feel like we have so few edits on each page that the user can just look the version history up for themselves and determine who contributed how much. If we want this for some kind of further processing by algorithms or just generally make it more suited for research it might be worth it.--SStark (talk) 20:58, 13 May 2024 (CEST)

Thanks for checking. A pity the tool is not supported for projects like ours. For us, the main goal would be transparency concerning authorship: who contributed how much to which article, who the main author or one of the main authors? At the moment, this is trivial - in almost all cases, the account initially creating an article is also the main author (and normally it's Linda), and to double-check a quick glance at the page history is sufficient. However, I hope this will change in the next years, and certainly things will get more messy in the long run. So a tool is not urgent at all, but I would be willing to invest some resources at some point. atb --Christof Rolker (talk) 12:20, 14 May 2024 (CEST)