BibTex escaping

[kosarko]

by @dan-zeman:
"ă" (SMALL A WITH BREVE) disappears from bibtex citation

[kosarko]

We should come up with a more sustainable solution, then listing all the mappings https://github.com/ufal/lindat-dspace/blob/lindat/dspace-oai/src/main/java/cz/cuni/mff/ufal/utils/BibtexUtil.java#L122
@amirkamran might know of a js library to do this for us

[stranak]

This might be a good moment to think about a more comprehensive citation solution that can generate not only bibtex, but also other formats and has support of styles.

Have a look at citeproc and where it is used (CrossRef, Open Office and many more places). There are also many implementations, JS, Java, Python, …

[kosarko]

Haven't had the time to actually inspect what citeproc/CSL is. Just found out there was some debate regarding this in dspace jira. https://jira.duraspace.org/browse/DS-1359 and there seems to be a patch of some sort in https://jira.duraspace.org/browse/DS-1224. Only there seems to be some issues/concerns regarding licensing

[kosarko]

I'm testing citeproc-java and the thing is, at the first glance, it does not escape as I'd expect. There might be a different bibtex style that does that and I just haven't found it yet.

edit: sample output

<div class="csl-entry"> @misc{Košarko_2016, title={Test dataset}, publisher={Some rather lenghty name of a publisher that really does not exist just for convenience to see it working}, author={Košarko, Ondřej}, year={2016}}</div>

but I'd expect Ko{\v s}arko, Ond{\v r}ej

Maybe we are overdoing it. Is there really no support for UTF-8 in bibtex? @vidiecan I think you've linked somewhere saying that the escapes are really the right way...

[kosarko]

There are some hints in http://tex.stackexchange.com/a/57827 (maybe that's about biblatex and not bibtex). Tried the book from there and the misc generated above in ufal's sharelatex...that's probably also not "vanilla" bibtex

\documentclass{article}
\usepackage[utf8]{inputenc}
...
\usepackage{natbib}
\usepackage{graphicx}
...
\bibliographystyle{plain}
\bibliography{references}

and the accented chars are displayed normally. The only trouble was with the accent in the generated key above (Košarko_2016)

What do we really want to support in the exports?

[stranak]

I am not sure that Bibtex even supports accents in keys. I would just generate different ASCII keys. Accents in the visible text are important, though.

-Pavel

On 08 Jun 2016, at 10:50, Ondřej Košarko [email protected] wrote:

There are some hints in http://tex.stackexchange.com/a/57827 (maybe that's about biblatex and not bibtex). Tried the book from there and the misc generated above in ufal's sharelatex...that's probably also not "vanilla" bibtex

\documentclass{article}
\usepackage[utf8]{inputenc}
...
\usepackage{natbib}
\usepackage{graphicx}
...
\bibliographystyle{plain}
\bibliography{references}

and the accented chars are displayed normally. The only trouble was with the accent in the generated key above (Košarko_2016)

What do we really want to support in the exports?

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.

[kosarko]

I am not sure that Bibtex even supports accents in keys. I would just generate different ASCII keys. Accents in the visible text are important, though.

My point was rather do we really need to "escape" Košarko as Ko{\v s}arko in the authors field?

[stranak]

On 08 Jun 2016, at 11:07, Ondřej Košarko [email protected] wrote:

I am not sure that Bibtex even supports accents in keys. I would just generate different ASCII keys. Accents in the visible text are important, though.

My point was rather do we really need to "escape" Košarko as Ko{\v s}arko in the authors field?

I think we do, since very few people use Biblatex and Bibtex, which everybody uses for latexing, doesn't support UTF-8, AFAIK.

-Pavel

[stranak]

On 08 Jun 2016, at 11:09, Pavel Stranak [email protected] wrote:

On 08 Jun 2016, at 11:07, Ondřej Košarko [email protected] wrote:

I am not sure that Bibtex even supports accents in keys. I would just generate different ASCII keys. Accents in the visible text are important, though.

My point was rather do we really need to "escape" Košarko as Ko{\v s}arko in the authors field?

I think we do, since very few people use Biblatex and Bibtex, which everybody uses for latexing, doesn't support UTF-8, AFAIK.

Clarification: Some implementations of bibtex now probably do support some accented characters, but unless we verify that most or all support at least latin1 + latin2 accents, we do need to escape.

Š is also in Finnish and Estonian, so could it be working because it is in Latin 1?

-Pavel

[dan-zeman]

I have had issues (repeatedly) with accented UTF-8 characters in .bib file not being displayed properly (actually not being displayed at all). Even when I don't have to escape the characters in the main text of the paper, where UTF-8 is properly recognized.

[stranak]

On 8. 6. 2016, at 13:34, Dan Zeman [email protected] wrote:

I have had issues (repeatedly) with accented UTF-8 characters in .bib file not being displayed properly (actually not being displayed at all). Even when I don't have to escape the characters in the main text of the paper, where UTF-8 is properly recognized.

The same had happened to me, I think I tried to google and then I was surprised when I found that Bibtex doesn't support UTF-8. It is several years back, but I think we should stil expect that on the users' side. LaTeX is not something that people tend to keep updated and using bleeding edge new stuff in.

Pavel

[kosarko]

In #568 I've added mappings from http://lindat.mff.cuni.cz/services/morph/code-latin2-table.html; hopefully all of them.

Lets' leave citeproc/csl for some other time #567