The issus for Persian language styles

[VahedCivil]

Hello I use Persian language for inserting citation and bibliography. The styling for Persian language has a little problem. The ( ) for English and Persian languages is different. The native MS Word inline citation doesn't have this issue. And also the chars , and ; in Persian is ، and ؛ The style used for below image is APA . And also the style used for below image is APA 6th . For APA 6th style used & instead of و The fixed style is: And another issue is different grammar with English and Persian languages. In English we said "6th Edition" but in Persian at first we called "Edition" and after com "6". For example: "ویرایش 6" Mendeley translate "Edition" to "ویرایش" correctly but the edition number must be come after "ویرایش"

[adam3smith]

I've moved this to the schema repository as it's unrelated to the visual editor.

[bwiernik]

So two problems to address:

1. Brackets/parentheses should be flipped when in right-to-left languages. That could be addressed with new punctuation terms like semicolon, or it could just be a processor instruction.
2. The question of edition 6 versus 6th edition. How should we address this? We could try to do accommodate using some form of logic on the ordinal terms, but I'm inclined to just say that this real change in formatting needs to be handled by making a new style with a default locale set.

@bdarcus @denismaier

[denismaier]

Re. 2. (ordinals): Haven't we had a similar case recently where we've introduced similar logic? I just can't remember where that was.

Re. 1: There are two more (sub-)problems here: a) Different engines differ in how well they handle RTL. Word, for example, is generally not an optimal choice. LibreOffice is much better here. b) Items where only the title is in a RTL language/script with the rest (author, publisher, etc.) being in Latin script pose additional challenges.

[bwiernik]

Re 2. Are you thinking of the placement of era terms in cs:date (AD vs CE)?

Re 1. a) I think that is something we don't need to worry about? That's up to the processor/application to figure out how to handle their target environments I think. b) Currently, we don't model the idea that an item has variables in multiple languages. If we did address that, I think the question would be what is the direction of the script the brackets are adjacent to. If the are adjacent to a LTR script, render LTR brackets. Otherwise, render RTL brackets (flipping them). Thinking about this, I am leaning toward making it a processor instruction for handling RTL scripts.

[bwiernik]

Regarding ordinals, edition is formatted in styles like this:

        <group delimiter=" ">
          <number variable="edition" form="ordinal"/>
          <label variable="edition" form="short"/>
        </group>

So, accommodating that would require some pretty special processing. I'm inclined to say that that needs to just be handled by writing a layout for that locale (e.g., in a new style).

[denismaier]

Re. Parentheses: another thing to keep in mind: in many (all?) RTL languages, numerals are nevertheless written LTR. This might be the source of some problems in some systems.

[bwiernik]

@fbennett Have you wrestled with the parentheses issue in RTL scripts at all?