Citations: non-page locators placed in front of citation

Citations: non-page locators placed in front of citation

[M. ‘quintus’ Gülker]

Dear all,

apologies for my frequent e-mails. It’s just that I am evaluating the
citations facility for me.

This time it’s about non-page locators. Take the following document:

    #+TITLE: Test
    #+AUTHOR: testauthor
    #+LANGUAGE: de
    #+bibliography: /tmp/mwe/mwe.bib
    #+cite_export: csl /tmp/mwe/juristische-schulung.csl

    Das ist ein Test [cite:@saenger2013gsr § 12 Rn. 488].

juristische-schulung.csl is
https://github.com/citation-style-language/styles/blob/e22b8a566bad9b4c7f52720f60dd875057a5d210/juristische-schulung.csl.

This is mwe.bib:

    @Book{saenger2013gsr,
      author       = {Ingo Saenger},
      title        = {Gesellschaftsrecht},
      year      = {2013},
      edition   = {2},
      publisher = {Franz Vahlen},
      location  = {München},
      langid    = {ngerman}}

Note how this work is not cited by page, but instead (which is common
among German judicial literature) by section number (§) plus margin number
(Rn.). Exporting this e.g. to HTML yields in Footnote 1:

    §  Saenger, Gesellschaftsrecht, 2. Aufl. (2013), 12 Rn. 488

That is rather unexpected. It has pulled the § sign in front of the
citation. The citation should have looked like this:

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), § 12 Rn. 488

If I replace the citation in the document with

    Das ist ein Test [cite:@saenger2013gsr p. 245].

then it produces the expected

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), 245

(the page locator label, "S." in German, is intentionally suppressed by
the CSL style, so this is not a bug. However, the § locator has to
appear and is not suppressed by the CSL style).

If I export with the default "bare" processor, the citation comes out
fine as "(Ingo Saenger, 2013 § 12 Rn. 488)".

I conclude from this that org somehow mistreats the locator if it is not
a page number. I used org 9.5 from MELPA and Citeproc.el from commit
34e66583d95a8d80fb5b9f2960f3382ca0e6d3ab.

Is it a bug or (again) my error?

  -quintus

-- 
Dipl.-Jur. M. Gülker | https://mg.guelker.eu | PGP: Siehe Webseite
Passau, Deutschland  | kontakt@guelker.eu    | O<

Re: Citations: non-page locators placed in front of citation

[Nicolas Goaziou]

Hello,

M. ‘quintus’ Gülker <post+orgmodeml@guelker.eu> writes:

> apologies for my frequent e-mails. It’s just that I am evaluating the
> citations facility for me.

On the contrary, feedback on citations is very much welcome. This is
a new features, and as such, has some rough edges.

> This time it’s about non-page locators. Take the following document:
>
>     #+TITLE: Test
>     #+AUTHOR: testauthor
>
>     #+LANGUAGE: de
>     #+bibliography: /tmp/mwe/mwe.bib
>
>     #+cite_export: csl /tmp/mwe/juristische-schulung.csl
>
>     Das ist ein Test [cite:@saenger2013gsr § 12 Rn. 488].
>
> juristische-schulung.csl is
> https://github.com/citation-style-language/styles/blob/e22b8a566bad9b4c7f52720f60dd875057a5d210/juristische-schulung.csl.
>
> This is mwe.bib:
>
>     @Book{saenger2013gsr,
>       author       = {Ingo Saenger},
>       title        = {Gesellschaftsrecht},
>       year      = {2013},
>       edition   = {2},
>       publisher = {Franz Vahlen},
>       location  = {München},
>       langid    = {ngerman}}
>
> Note how this work is not cited by page, but instead (which is common
> among German judicial literature) by section number (§) plus margin number
> (Rn.). Exporting this e.g. to HTML yields in Footnote 1:
>
>     §  Saenger, Gesellschaftsrecht, 2. Aufl. (2013), 12 Rn. 488
>
> That is rather unexpected. It has pulled the § sign in front of the
> citation. The citation should have looked like this:
>
>     Saenger, Gesellschaftsrecht, 2. Aufl. (2013), § 12 Rn. 488

[...]

> Is it a bug or (again) my error?

It is a bug. You use a non-breaking space between the locator and the
number. I hadn't anticipated this (duh!). I fixed it. Could you confirm
it?

Thank you.

Regards,
-- 
Nicolas Goaziou

Re: Citations: non-page locators placed in front of citation

[M. ‘quintus’ Gülker]

Am Sonntag, dem 10. Oktober 2021 schrieb Nicolas Goaziou:
> On the contrary, feedback on citations is very much welcome. This is
> a new features, and as such, has some rough edges.

Thanks for bearing with me.

> It is a bug. You use a non-breaking space between the locator and the
> number. I hadn't anticipated this (duh!). I fixed it. Could you confirm
> it?

Actually, I have remapped the § key to automatically always insert an
NBSP following it -- I need the § sign all the time and in normal text
line-breaking after the § violates typographic rules. Thus I did not
notice I had an NSBP there. Apologies for not mentioning this.

I however do not think the problem is related to the NBSP. I just
retried without it, and the § sign is still pulled towards the front.
I also retried with current Git (Org mode version 9.5
(release_9.5-93-gd87250 @ /home/quintus/.emacs.d/org-mode/lisp/)), but
that does not change anything. Either with the NBSP or with a normal
space, the § sign is pulled toward the front. That is,

    Das ist ein Test [cite:@saenger2013gsr § 12 Rn. 488].

Still yields:

    §  Saenger, Gesellschaftsrecht, 2. Aufl. 2013, 12 Rn. 488

Without the NBSP, that is,

    Das ist ein Test [cite:@saenger2013gsr § 12 Rn. 488].

it yields:

    § Saenger, Gesellschaftsrecht, 2. Aufl. 2013, 12 Rn. 488

Neither is correct; the § should go after the "2013, ".

  -quintus

-- 
Dipl.-Jur. M. Gülker | https://mg.guelker.eu | PGP: Siehe Webseite
Passau, Deutschland  | kontakt@guelker.eu    | O<

Re: Citations: non-page locators placed in front of citation

[Nicolas Goaziou]

Hello,

M. ‘quintus’ Gülker <post+orgmodeml@guelker.eu> writes:

> I however do not think the problem is related to the NBSP. I just
> retried without it, and the § sign is still pulled towards the front.
> I also retried with current Git (Org mode version 9.5
> (release_9.5-93-gd87250 @ /home/quintus/.emacs.d/org-mode/lisp/)), but
> that does not change anything. Either with the NBSP or with a normal
> space, the § sign is pulled toward the front. That is,
>
>     Das ist ein Test [cite:@saenger2013gsr § 12 Rn. 488].
>
> Still yields:
>
>     §  Saenger, Gesellschaftsrecht, 2. Aufl. 2013, 12 Rn. 488
>
> Without the NBSP, that is,
>
>     Das ist ein Test [cite:@saenger2013gsr § 12 Rn. 488].
>
> it yields:
>
>     § Saenger, Gesellschaftsrecht, 2. Aufl. 2013, 12 Rn. 488
>
> Neither is correct; the § should go after the "2013, ".

Then, this may be a bug in Citeproc library itself. I suggest to report
it upstream.

Regards,
-- 
Nicolas Goaziou

Re: Citations: non-page locators placed in front of citation

[M. ‘quintus’ Gülker]

Am Sonntag, dem 10. Oktober 2021 schrieb Nicolas Goaziou:
> Then, this may be a bug in Citeproc library itself. I suggest to report
> it upstream.

Done: https://github.com/andras-simonyi/citeproc-el/issues/57

  -quintus

-- 
Dipl.-Jur. M. Gülker | https://mg.guelker.eu | PGP: Siehe Webseite
Passau, Deutschland  | kontakt@guelker.eu    | O<

Re: Citations: non-page locators placed in front of citation

[András Simonyi]

Dear All,

On Sun, 10 Oct 2021 at 22:11, Nicolas Goaziou <mail@nicolasgoaziou.fr> wrote:
>
> Then, this may be a bug in Citeproc library itself. I suggest to report
> it upstream.

looks like an Org (oc-csl) side locator parsing problem to me, because
using the alternative [cite:@saenger2013gsr para. 12 Rn. 488] form I
seem to get the correct result. Can it be a regex matching problem
with the paragraph symbols?

best wishes,
András

Re: Citations: non-page locators placed in front of citation

[M. ‘quintus’ Gülker]

Am Sonntag, dem 10. Oktober 2021 schrieb András Simonyi:
> looks like an Org (oc-csl) side locator parsing problem to me, because
> using the alternative [cite:@saenger2013gsr para. 12 Rn. 488] form I
> seem to get the correct result. Can it be a regex matching problem
> with the paragraph symbols?

Now it’s getting wild. If I use

    Das ist ein Test [cite:@saenger2013gsr para. 12 Rn. 488].

Then it exports to this:

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), ¶¶ 12 Rn. 488

The locator’s position is correct, but it now has two pilcrow signs ¶
instead of one section sign §. I have never seen that in citations, but
maybe it is actually correct for other disciplines than mine.

FWIW, if I use

    Das ist ein Test [cite:@saenger2013gsr line 12 Rn. 488].

it yields the correct

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), Z. 12 Rn. 488

(with "Z." being short for "Zeile", i.e. "line" in German).

So this appears to be a problem specifically with the § sign?

  -quintus

-- 
Dipl.-Jur. M. Gülker | https://mg.guelker.eu | PGP: Siehe Webseite
Passau, Deutschland  | kontakt@guelker.eu    | O<

Re: Citations: non-page locators placed in front of citation

[Nicolas Goaziou]

Hello,

M. ‘quintus’ Gülker <post+orgmodeml@guelker.eu> writes:

> Now it’s getting wild.

Indeed. The "fix" I introduced was a mistake. I pushed a new fix.
I think the initial issue is solved now. Could you confirm it?

Thank you.

Regards,
-- 
Nicolas Goaziou

Re: Citations: non-page locators placed in front of citation

[M. ‘quintus’ Gülker]

Am Montag, dem 11. Oktober 2021 schrieb Nicolas Goaziou:
> Indeed. The "fix" I introduced was a mistake. I pushed a new fix.
> I think the initial issue is solved now. Could you confirm it?

Thank you so far -- we are getting closer. Now at »Org mode version 9.5
(release_9.5-102-gd0b557 @ /home/quintus/.emacs.d/org-mode/lisp/)«, the
/positioning/ is correct. What is still not correct is the character
used for the locator label. For me,

    Das ist ein Test [cite:@saenger2013gsr § 12 Rn. 488].

now (regardless of a NSBP or normal space) exports to:

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), ¶¶ 12 Rn. 488

As explained, the positioning is now correct. However, instead of the
section § sign it now has two pilcrow signs ¶¶. It should be just one
sign, and it should be a section sign rather than a pilcrow one.

  -quintus

-- 
Dipl.-Jur. M. Gülker | https://mg.guelker.eu | PGP: Siehe Webseite
Passau, Deutschland  | kontakt@guelker.eu    | O<

Re: Citations: non-page locators placed in front of citation

[Bruce D'Arcus]

On Mon, Oct 11, 2021 at 1:57 PM M. ‘quintus’ Gülker
<post+orgmodeml@guelker.eu> wrote:

> As explained, the positioning is now correct. However, instead of the
> section § sign it now has two pilcrow signs ¶¶. It should be just one
> sign, and it should be a section sign rather than a pilcrow one.

Looks like § is currently mapped to the same as ¶: "paragraph", which
is indeed incorrect.

https://github.com/citation-style-language/locales/blob/0cc3885f6100e26ac6c6d103efa6f3d7195fd21b/locales-de-DE.xml#L210

Bruce

Re: Citations: non-page locators placed in front of citation

[M. ‘quintus’ Gülker]

Am Montag, dem 11. Oktober 2021 schrieb Bruce D'Arcus:
> Looks like § is currently mapped to the same as ¶: "paragraph", which
> is indeed incorrect.
>
> https://github.com/citation-style-language/locales/blob/0cc3885f6100e26ac6c6d103efa6f3d7195fd21b/locales-de-DE.xml#L210

This is interesting. Pandoc -- which I thought relies on these files? --
outputs the expected § sign. For

    Das ist ein Test [@saenger2013gsr, § 12 Rn. 488].

it yields

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), § 12 Rn. 488↩

which is what I would have expected. The issue appears to be in the way
§ is interpreted in the text input.

After reading the locales-de-DE.xml file, I suspect a case of lost in
translation. In German, when read out aloud, we spell the “§” character
as “paragraph”. This is evidently different from the English speaker,
who, as I have been told, would read it out as “section”. However, in
fact the German word “Paragraph” translates to “section” in English. The
unit below it is called “Absatz” in German -- in English, it would be
“paragraph”. That is:

| English   | German    | German Short form |
|-----------+-----------+-------------------|
| Section   | Paragraph | §                 |
| Paragraph | Absatz    | Abs.              |

At least, when it comes to legal contexts. As it always is the case with
language, meaning can vary be context. If I talk about sections of a
novel, in German it would be “Abschnitt” rather than “Paragraph”. But
that’s not something one would cite. If I really would need to cite a
novel, I would cite it either by chapter or by page number (with line
number, if necessary).

There is no symbolic sign to my knowledge to denote the unit below § in
German. The locales-de-DE.xml files suggests ¶. That might actually be
valid in some discipline, but we do not use that in jurisprudence. I do
not see a problem with it; it can be left in. What needs to change is
the part of org that maps § to CSL's "paragraph". § needs to be mapped
to "section" instead.

Long story short: I do not think that it is a bug in locales-de-DE.xml,
and I guess Pandoc proves my point here. Please map § to "section"
instead of "paragraph" in org-cite, i.e., do it the way Pandoc does it.

  -quintus


-- 
Dipl.-Jur. M. Gülker | https://mg.guelker.eu | PGP: Siehe Webseite
Passau, Deutschland  | kontakt@guelker.eu    | O<

Re: Citations: non-page locators placed in front of citation

[Nicolas Goaziou]

M. ‘quintus’ Gülker <post+orgmodeml@guelker.eu> writes:

> Long story short: I do not think that it is a bug in locales-de-DE.xml,
> and I guess Pandoc proves my point here. Please map § to "section"
> instead of "paragraph" in org-cite, i.e., do it the way Pandoc does
> it.

I mapped both § and §§ to "section". Hopefully, the issue is now
completely fixed.

Regards,

Re: Citations: non-page locators placed in front of citation

[M. ‘quintus’ Gülker]

Am Dienstag, dem 12. Oktober 2021 schrieb Nicolas Goaziou:
> I mapped both § and §§ to "section". Hopefully, the issue is now
> completely fixed.

Nearly :-). Thank you so much for your hard work so far! It is
incredible to see how this develops.

With »Org mode version 9.5 (release_9.5-104-g2b1fc6 @ /home/quintus/.emacs.d/org-mode/lisp/)«
now

    Das ist ein Test [cite:@saenger2013gsr § 12 Rn. 488].

gives:

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), §§ 12 Rn. 488

Positioning and choice of locator label is correct. The one thing left
is that it uses double §§ instead of a single §, that is, it treats the
citation as a plural one whereas it should be a singular one. The same
happens with

    Das ist ein Test [cite:@saenger2013gsr section 12 Rn. 488].

Interestingly, the input

    Das ist ein Test [cite:@saenger2013gsr § 12].

(which has no suffix) does export to

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), § 12

with only one § sign. Then, a real multi-section citation like this:

    Das ist ein Test [cite:@saenger2013gsr §§ 12 ff.].

instead yields a single §:

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), § 12 ff.

This one however:

    Das ist ein Test [cite:@saenger2013gsr §§ 12-14].

is correct again:

    Saenger, Gesellschaftsrecht, 2. Aufl. (2013), §§ 12-14

This is a little confusing. Maybe an error on my part? Do I need to
signal singular and plural use more explicitely?

  -quintus

-- 
Dipl.-Jur. M. Gülker | https://mg.guelker.eu | PGP: Siehe Webseite
Passau, Deutschland  | kontakt@guelker.eu    | O<

Re: Citations: non-page locators placed in front of citation

[András Simonyi]

Dear All,

On Tue, 12 Oct 2021 at 07:43, M. ‘quintus’ Gülker
<post+orgmodeml@guelker.eu> wrote:

>     Das ist ein Test [cite:@saenger2013gsr § 12 Rn. 488].
>
> gives:
>
>     Saenger, Gesellschaftsrecht, 2. Aufl. (2013), §§ 12 Rn. 488
[...]
> This is a little confusing. Maybe an error on my part? Do I need to
> signal singular and plural use more explicitely?

No, it's citeproc-el which is supposed to determine whether the passed
locator string represents single vs multiple references, using § vs §§
shouldn't make a difference, because they simply signal that the label
is "section". Apparently, citeproc-el's classifier algorithm doesn't
handle some of your examples correctly -- could you open an issue
about this on the project page? Thanks in advance!

best wishes,
András

Re: Citations: non-page locators placed in front of citation

[M. ‘quintus’ Gülker]

Am Dienstag, dem 12. Oktober 2021 schrieb András Simonyi:
> No, it's citeproc-el which is supposed to determine whether the passed
> locator string represents single vs multiple references, using § vs §§
> shouldn't make a difference, because they simply signal that the label
> is "section".

I see. This explains why it makes no difference whether I use "section"
or "§". Trying to find this out from the subsequent text alone sounds
ambitious to me, though. There could be difficult edge cases I imagine.

> Apparently, citeproc-el's classifier algorithm doesn't
> handle some of your examples correctly -- could you open an issue
> about this on the project page? Thanks in advance!

Done: https://github.com/andras-simonyi/citeproc-el/issues/58

  -quintus


-- 
Dipl.-Jur. M. Gülker | https://mg.guelker.eu | PGP: Siehe Webseite
Passau, Deutschland  | kontakt@guelker.eu    | O<

Re: Citations: non-page locators placed in front of citation

[Nicolas Goaziou]

Hello,

András Simonyi <andras.simonyi@gmail.com> writes:

> looks like an Org (oc-csl) side locator parsing problem to me, because
> using the alternative [cite:@saenger2013gsr para. 12 Rn. 488] form I
> seem to get the correct result. Can it be a regex matching problem
> with the paragraph symbols?

You're right. I thought I had solved the regexp part but I was wrong.
I pushed a new attempt to solve this.

Sorry for the noise.

Regards,
-- 
Nicolas Goaziou