translating word keys containing the word "in"

[Pavel Chernitsky]

Firstly, apologies for any typos or for using the wrong names for things (you'll see what I mean in sec). English isn't my first language.

OK, let's get to it.

there is an issue I came across that affects translations and I would like to hear from you guys, if any of you came across something similar. 

We use Hebrew for our community site and we're doing the translating by ourselves using only the built in translator (not modifying the HTML/CSS/whatever else affects that kind of thing, I'm no programmer so I have no idea what it's called). 

There are some phrases that are build like so (I added the brackets here for reading domfort): [user_other_activity_reply] [Forums_forum_1].
The resulting text comes out as: [%2$s replied to %3$s's %1$s in] [Forum_1 name] 
Those are TWO keys - separated by a space - remember that space, it's gonna come back
In normal human speak, on the front end, you'll see: "Mike replied to Pavel Chernitsky's topic in feedback and ideas"

This, of course, make perfect sense for english and other languages that use separate words as a preposition (and thanks to Google for teaching me what it's called in English 😄). However, in languages that just use a single letter that is prefixed to the word without a space (resulting in - very roughly - a word like "inforum") like Hebrew (that I know of, pretty sure are are others) the resulting phrase comes out looking kinda dumb. Basically, the issue is that you cannot remove the space (told you it'll come back) between the "in" and the thing (forum/club/gallery/etc).

I can think two fixes for that:
1. moving the "in" from the "user_other_activity_reply" key (%2$s replied to %3$s's %1$s in) to the "Forums_forum_1" key (Forum_1 name), thus resulting in making the admin able to decide whether or not they want to have a space between "in" and "forum" or not. 
2. Adding a space to the end of the "user_other_activity_reply" key and making it's deletion possible.

 

Did any of you come across such an issue? Can you think of another, maybe simpler, fix for that? And also, thanks for anyone who has read this far, I imagine it couldn't have been easy 🤣

Edited July 10, 2020 by Pavel Chernitsky

[MMXII]

It is a huge problem that many phrases consist of language string fragments. While this works in English, it is a serious problem for many other languages.

Sadly the translation section of the software has not seen many improvements in the past (and I highly doubt it will at all in the 4.x lifecycle). I hope that for version 5.x IPS talks to some foreign language speakers to get some feedback before a new translation system will be created.

[bfarber]

We actually made massive improvements to the language system in 4.x (such as the pluralization system we have in place now), and we introduced some new improvements to managing language strings for 4.5. This sort of thing is something we can address when it's reported. It might be best if you made a list of any language strings where you run into problems and we can address them on a case by case basis.

[Pavel Chernitsky]

Most of the things I came across are solvable, some things require more tinkering or creativity (or just knowing coding, I don't, so maybe some of those things are actually really simple) but after understanding the "how" of the translation system can be solved or worked around rather easily.

An example for a thing I'm having the a lot of unsolvable trouble is the fact that english is a very "efficient" language in regards to pronouns. in relation to hebrew, the ratio is 8:1 - meaning there are 8 pronouns in Hebrew for every one in English.

Two examples of this are:
1. "You" in english refers to four groups of people: singular male, singular female, plural male and plural female. Hebrew has a word for each of those. 
2. English doesn't have a distinction between male and female verbs - "He reacted, She reacted, It reacted" as well. Hebrew has different variants for male and female, and none for the "Its" - who are treated as one or the other (Male or female). 

If I'm totally honest, I don't think the translation system needs that much work. what would be of massive help is a "how it works - tips and tricks" resource with diving deep into things like "where does the system take words from" (the %s thing) or the different customization options of, say, the pluralization system (like using {!#[1=message one][?=# messages]}" instead of {#[1=message][?=messages]} for languages the place the number of the item in different places depending on the number. 

If you guys want, I'm more than willing to work with you guys more closely on the translation system, or consult on the the relations between it and English, where they are similar and where they differ.

[MMXII]

2 hours ago, bfarber said:

We actually made massive improvements to the language system in 4.x (such as the pluralization system we have in place now)

If you are referring to my statement: I was writing about improvements within the 4.x releases. Pluralization is great. The same goes for the possibility to swap positioning of %s placeholders. And you are correct that with 4.5 there are some improvements regarding language string handling.

But regarding translating the software, at least for me, the real problems lie somewhere else. Fractured language strings are just an example. Those cause so many problems: grammar-wise, reference-wise...

Two more have just been mentioned above:

1 hour ago, Pavel Chernitsky said:

Two examples of this are:
1. "You" in english refers to four groups of people: singular male, singular female, plural male and plural female. Hebrew has a word for each of those. 
2. English doesn't have a distinction between male and female verbs - "He reacted, She reacted, It reacted" as well. Hebrew has different variants for male and female, and none for the "Its" - who are treated as one or the other (Male or female). 

Then there is the usage of "%s's". Sometimes changing it to "... of %s" does work. Usually it does not.

I am sure a flexible translation system is a huge undertaking. So all I am saying is that if there are plans to improve the system in a future release, talking to foreign language speakers in the first place will be very valuable to get feedback regarding much needed improvements.