5

I use biblatex and biber. I used to use bibtex. So, when I first started writing .bib files I would have entries like this:

@article{descartes,
    Author = {Descartes, Ren{\'{e}}}
    Title = {`Something' --- Something}
}

Now I do this:

@article{descartes,
    Author = {Descartes, René}
    Title = {`Something' --- Something}
}

I use the unicode accented e, but I don't replace the single quotes or dash with unicode equivalents. This works fine, but it seems to me that I'm probably mixing two conventions for no reason. (I also use two hyphens rather than one for a page range, even though I know that I don't need to with biblatex.)

Is there any specific reason why it would be best practice to change to using unicode for everything, or any reason why it wouldn't.

  • 4
    I would use \enquote for the quotes so that can be adapted in the document. – Ulrike Fischer Mar 29 '15 at 20:20
  • 2
    Or \mkbibquote, which handles the annoyances of the "American" rules for quotation marks and trailing punctuation. (I forget if biblatex will do that with csquote's \enquote command.) – jon Mar 29 '15 at 20:43
  • 2
    You are not really mixing 2 conventions except superficially. Accented characters are one thing. Ligatures are another. At least for standard TeX/pdfTeX, -- and --- are ligatures in the same way that fi and ffi etc. are. So, even if consistency were a particular virtue here, you wouldn't need to replace -- and --- unless you also input the fi etc. ligatures directly. (And I don't know why you'd do that.) If you use XeTeX/LuaTeX, things might be a bit different.... – cfr Mar 29 '15 at 22:28
  • 2
    Look at the csquotes documentation. For example, I use \MakeAutoQuote{‘}{’} and then use ‘something’. – cfr Mar 29 '15 at 23:20
  • 2
    I think it is sensible to enter UTF8 glyphs that can easily be produced with your keyboard directly (è, a, ß - of course that all depends on the keyboard layout, and portability might suffer) while characters that are harder to produce (en-dash, em-dash) (not knowing their UTF8 code point) can be entered with their macro. It's good to weigh a neat and easily readable picture in the file against ease of inoput. In some fonts or editors it might be quite hard to distinguish certain glyphs (e.g. b-b–b—b/b-b–b—b), in which case the macro greatly improves readability. – moewe Mar 30 '15 at 6:10
3

There is no technical reason not to mix utf8 text and LaTeX in one .bib file. Biblatex, with biber, handles both fine. Ease of input, readability, and portability may count in favour of various approaches.

  • For some fonts, especially fixed width ones, hyphens, en-dashes, and em-dashes are hard to distinguish.
  • Some characters are hard to input, especially dashes. Others, depending on the keyboard, are easier.
  • If the quotes are done as `, ', ``, and '' (i.e. LaTeX) then they won't adapt to the quote settings controlled by the document's language. Using \enquote allows them to be, but might be considered less readable. Using csquote.sty's active characters is a better solution e.g. \MakeAutoQuote{‘}{’}.

If you use BibDesk it's also worth thinking about exporting files. Using export templates it's possible to convert LaTeX or utf8 plain text to rich text. It's also possible to convert LaTeX dashes to unicode, but not vice versa. It's also not possible to convert LaTeX quotes to smart unicode quotes.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.