citeproc - en-dash/hyphen in bibliography - regression?

adamsmith · February 9, 2012

If I understand correctly, we agreed and implemented that Zotero would convert all hyphens in page-ranges to en-dashes.
(e.g. here http://forums.zotero.org/discussion/1754/hyphenen-dash-and-ppp/ )
That is the case for in-text citations, but Zotero doesn't do this (anymore?) in bibliographies (tried clipboard and LO plugin w. 3.0.1)

Am I missing something?

Rintze · February 10, 2012

If I remember correctly, we settled on the rather arbitrary rule that hyphen replacement only took effect when the page-range-format attribute was set. Is that the case for the style you're testing?

See
https://github.com/citation-style-language/documentation/blob/74cbf1fefe7c3826ce8752c8625a9f2f5120e9e3/specification.txt#L2043

adamsmith · February 10, 2012

that's crazy. Uhm. I guess I'll start putting pange-range-formats into styles then?

Rintze · February 10, 2012

I'm not very happy with it either. I think the main reason for this decision was that Frank argued we needed a way to display the contents of the "page" variable untouched, as for law citations the field is likely to contain more than just page ranges. Can't find the exact discussion. This isn't really an issue for styles for the sciences.

Apart from all that, having page-range-format set in styles is generally desirable, yes.

adamsmith · February 10, 2012

lawyers just want to be disliked, don't they? ;-)

fbennett · February 12, 2012

Yes, those were crazy days. My judgment was probably under strain from juggling code in my head. I've just checked in a processor version that applies en-dash to recognizable numeric ranges in all cases. It's a fairly conservative change, but should address the most common cases.

(I'm catching up on patches in MLZ at the moment. As soon as that's done, the updated processor will be available for merging to the official 3.0 branch.)

adamsmith · February 12, 2012

Thanks Frank!
that's the .279 build?

fbennett · February 12, 2012

Yep.

In the MLZ extended schema, I'm forcing page and locator to be handled as cs:number, which applies more robust logic (the string is split into units if possible, and affixes are applied to each unit, preserving the commas, ampersands and rangey connectors). It also recognizes overrides for en-dash (--) and hyphen (\-). The cs:text logic is more primitive, but should cover most cases. I waffled a little over the solution. If folks run into problems it can be adjusted, as ever.