Citations with biblatex -- locator information partly lost when using \autocite and german localization

[mpapesch]

Hi,

i just spent a couple hours figuring out what seems to be going on.

For some of my citations, parts of the locator information were dropped in the resulting document.

German locale specifics:

• "vgl." corresponds to "cf."
• "S." is short for "Seite"; which is "Page" in english
• "ff" is the equivalent to "pp." and implicitly means 'this page and the one following'

text [vgl. @test, S. 55 ff] which should render to text (vgl. test 1999, S. 55 ff) (which would translate to english: [cf. @test, pp. 55--56])

Without German locale, S. 55 ff is not recognized as locator information and copied verbatim -- yielding the expected result in the .pdf document:

host:/www# echo -e "text [vgl. @test, S. 55 ff] \n\ntext  [vgl. @test, {S. 55 ff}]" | pandoc --from=markdown --to=latex --biblatex
text \autocite[vgl.][S. 55 ff]{test}

text \autocite[vgl.][S. 55 ff]{test}

With German locale, however, S. 55 ffis recognized as locator information:

host:/www# echo -e "text [vgl. @test, S. 55 ff] \n\ntext  [vgl. @test, {S. 55 ff}]" | pandoc --from=markdown --to=latex --biblatex -V lang=de
text \autocite[vgl.][55 ff]{test}

text \autocite[vgl.][55 ff]{test}

Now, while pandoc does recognize S. 55 ff as locator information and strips the leading S. from it; biblatex does not recognize the remaining part 55 ff as locator information and, consequently, does not prepend the S. page(s) marker, yielding the incorrect result: text (vgl. test 1999, 55 ff) -- notice the missing "S.".

In the BibLaTeX documentation, I found that \DeclareRangeChars{~,;-+/} defines the characters which are acceptable for a range of pages in locator information. Adding the letter f to this macro \DeclareRangeChars{~,;-+/f} in the preamble of the created .tex document, fixed the problem for me and yields the correct result in the .pdf document: (vgl. test 1999, S. 55 ff)

Questions:

1. Is the observed behavior desired (or am I possibly doing something wrong here)?
2. Is there an easier way to protect the locator information, so that is used verbatim in the resulting \autocite command?

Thanks for any feedback.
Matthias

[jgm]

2 is a good question for someone with biblatex expertise...
If there is a way, we could adopt it.

[njbart]

@mpapesch - With LaTeX/BibLaTeX as the target format, you might consider replacing your ff with the native BibLaTeX command for the suffix ‘the following pages’, which is \psqq (for a single page, you can use \psq). However, you will need to insert a space before \psqq (contrary to BibLaTeX's recommendations), as pandoc will not allow it to pass through otherwise. I have not noticed any differences in spacing in the PDF output, though. When using \psqq, there is no need to redefine \DeclareRangeChars. Additionally, this approach ensures that the string is properly localised. For further information, see the BibLaTeX manual, version 3.20, page 119.

[mpapesch]

💡 ahhh - that sounds like a good solution! In fact, had some unfortunate results where the page number "55" and the "ff" were separated by a linebreak in a footnote citation. I was able to fix this problem by replacing the space with a (unicode) non-breaking space in the markdown source file.

Up to now, most my pandoc usage translates markdown to LaTeX (and finally pdf). In the near futre, I'd like to also create html and plain text versions of the same markdown source. I wonder, how would citeproc react to a citation "[vgl. @test, S. 55 \plqq]"

host:/www# echo -e "text [vgl. @test, S. 55 ff] \n\ntext  [vgl. @test, S. 55 \psqq]" | pandoc --from=markdown --to=plain --citeproc -V lang=de
[WARNING] Citeproc: citation test not found
text (vgl. test?, S. 55 ff)

text (vgl. test?, S. 55 )

Would it be possible to adapt the pandoc locator-parser for the german locale to translate a postfix 'f' to '\psq' and 'ff' to '\psqq'?

After all, the preceding 'S.' is already being removed correctly; and pandoc itself could follow the recommendation not to add a space after the page number.

If I am not mistaken, citeproc would just pass 'ff' on to plaintext output.

Does that make sense?