German translation for "Ed. by" etc. is wrong if editor is an institute or female

[nikonak]

I encountered a problem with the German translation for "byeditors" etc. For a book entrytype, setting the editor field gives "Hrsg. von (editor)". This is fine if the editor is an actual person.

But consider the following MWE:

\documentclass[ngerman]{article}
\usepackage[utf8]{inputenc}
\usepackage{babel}
\usepackage{biblatex}

\begin{filecontents}{\jobname.bib}
  @book{Muster_2020,
    author = {Anna Musterfrau},
    gender = {sf},
    editor = {{Muster-Institut}},
    title = {Musterhafte Studie},
    langid = {german},
    location = {Musterstadt},
    year = {2020}
  }
\end{filecontents}
\addbibresource{\jobname.bib}

\begin{document}
  \nocite{Muster_2020}
  \printbibliography
\end{document}

Here, editor is not an actual person, but an institute. As such, it needs an article (emphasised below).

In the above example, the institutes gender is neutral (and it would be the same for masculine), so it should print "Hrsg. von dem Muster-Institut" or "Hrsg. vom Muster-Institut". If the institute had a feminine gender, it should print e.g. "Hrsg. von der Muster-Anstalt".

If a verbose style is used, the same is true for an actual female person: "Anna Musterfrau, Herausgeber" is printed instead of "Anna Musterfrau, Herausgeberin". This also applies to the (admittedly seldomly used, at least in my case) translations for compiler, redactor etc.

Now, I understand that this would require not only expanding the datamodel by e.g. editor(a,b,c)isnonperson and editor(a,b,c)gender (or only editor(a,b,c)gender with distinct fields for people and institute, e.g. sm/ism, sn/isn etc.), but also rewriting lots of the current way biblatex handles editors. This is by no means a feature request, but a call to attention as I could imagine the same problem persists in other languages. For anyone encountering this problem, two possible workarounds are:

1. Changing the translations to not give "Hrsg. von (editor)" etc., but "Hrsg: (editor)", "(editor), Hrsg." (similar as for article, incollection etc.); "Hrsg. v. (editor)" doesn't solve the feminine case, unfortunately.
2. Using the titleaddon field to print the desired string verbatim (i.e. in above case setting titleaddon to "Hrsg. vom Muster-Institut" while unsetting editor).

EDIT¹: Found the following third workaround that I think is the most sensible thing to do at the moment:

3. Setting editortype to the desired verbatim string will also print the desired result, i.e. setting editortype to "Hrsg. vom" will print "Hrsg. vom Muster-Institut".

[1:] Also deleted some private information from MWE and deleted edit history.

[moewew]

The "Evangelische Kirche Deutschland" example is actually really tricky, because you'd want

Evangelische Kirche Deutschland (Hrsg.): ...

but

Hrsg. von der Evangelischen Kirche Deutschland

from the same input depending on the context or style preferences. So this not only requires support from bibstrings and bibmacros, but also different inputs (since I don't think we want to try to have biblatex decline field contents).

[nikonak]

God forbid! I think that example really is an edge case. It has an academically accepted shorthand (EKD) anyway. But then again, institutes with adjectives in their name are not uncommon at all.

Maybe multiscript support could be facilitated for different inputs? Food for thought.

[gusbrs]

For the record, I have a similar case for letters which might be worth mentioning here.

A typical letter reference/citation will start with "John Doe to James Smith", but "John Doe to the Ministry XYZ" if it is some sort of organization/entity. Depending on treatment, a person might also deserve an article: "John Doe to the Director XYZ".

I've dealt with it by creating a recipientgender field, constrained to sf|sm|sn|pf|pm|pn and by having corresponding bibstrings in the languages of interest. I could likely have done something similar with field annotation. But, either way, the style has to support it.

Just another data point.

[moewew]

Maybe multiscript support could be facilitated for different inputs? Food for thought.

Yeah, I think that could work. The dative case could just be a 'version' of the field. Of course it requires user intervention (and some doubling of the input author = {{Evangelische Kirche Deutschland}}, author_dat = {{der Evangelischen Kirche Deutschland}},), but then it would probably be fairly straightforward.

With field annotations available it would probably be more natural to give the gender and other necessary information 'institution or not' in a field annotation. (The current gender field is a bit unfortunate, since it can apply to author or editor depending on field availability, but not separately to each of them.) Then this could even be done per item if desirable.