languages for langid and for languages field differ and look pretty random

[johanneswilm]

These fields correspond to predefined values that exist for either field. So maybe there is not a lot we can do about this, but it will seem odd to any end user. I think it would be interesting to hear why biblatex doesn't add translation strings for some more of these, for example.

For language, these are defined by the translation strings that are predefined by biblatex:

['catalan', 'croatian', 'czech', 'danish',
'dutch', 'english', 'american', 'finnish', 'french', 'german', 'greek',
'italian', 'latin', 'norwegian', 'polish', 'portuguese', 'brazilian', 'russian',
'slovene', 'spanish', 'swedish']

(Note the absence of Chinese, among others, and the odd labeling for US English: "american").

Any custom value will be accepted in the language field, but these will be printed outright: Only the above will be translated to other languages automatically.

For langid the options are defined through support from babel/polyglossia and we currently have:

["acadian", "afrikaans", "arabic", "basque", "bulgarian", "catalan", "pinyin", "croatian", "czech", "danish", "dutch", "australian", "canadian", "newzealand", "ukenglish", "usenglish", "estonian", "finnish", "french", "canadien", "ngerman", "naustrian", "greek", "hebrew", "hungarian", "icelandic", "italian", "japanese", "latin", "latvian", "lithuanian", "magyar", "mongolian", "norsk", "nynorsk", "farsi", "polish", "portuguese", "brazilian", "romanian", "russian", "serbian", "serbianc", "slovak", "slovene", "spanish", "swedish", "thai", "turkish", "ukrainian", "vietnamese"]

We also support various aliases of these. I assume that there is an actual language difference between Cyrillic Serbian and Serbian, and not just two different packages. The langid field does not allow custom input, but the above list covers all our test files.