More variables for czech translators

[Aladalar]

Hi there,
We wont in our translate use more than premade translation. Grammar structure in our language is totaly different than Englis and we have lack of variables in backstrory and others.

In Keyed/Grammar is defined these variables HIS,HISCAP,HIM,HIMCAP,HE,HECAP
but in our translate we will need more than this, at least two more.

Is there any possibility how moke more then theese?

[Tynan]

Can you say specifically what you need?

[Aladalar]

I dont know how describe it but i you have theese HIS,HISCAP,HIM,HIMCAP,HE,HECAP we need expand them at least for two more to keep our language correctli and nice, or we will do some patches which will work but dont look good

[Tynan]

What two more do you need?

Can you give an example of the problem you need to solve, and what you want to do to solve it?

[Aladalar]

Eng have same verb for both gender but in our lang verb is changed with gender so we need modify them

for ex:
Male: Dělal práci
Female: Dělala práci

[Tynan]

Interesting, verbs have genders based on their subject? Interesting.

Unfortunately there are a lot of variations on this in different languages and we can't cover them all.

This bug describes them: https://ludeon.com/mantis/view.php?id=683

I hope we'll eventually get to address this, but it could be some time.

[Aladalar]

We just need to add some kind of custom variables for example:
<Prohe>HE</Prohe> -> this you alredy made
<CustomHe>HE</CustomHe> -> this you can made and usage will be anything after Custom
or for ex CustomFemHe/CustomMalHe for both gender

Or second idea where exactli in main code is enum of your variables so i can create dll with more variables which we are lack of?

[Tynan]

I still don't really understand how Czech works here.

Can you describe exactly where the gender comes from, what it applies to?

And, can you give 3 examples of sentences where this occurs, with translations, describing each word and what it does?

Thanks.

[Aladalar]

Ok, Ill contact translate team and ask them, but it can take long and still we dont have patch until new version is there some kind of my patch like in mod (assemblies/czech.dll)

[Aladalar]

So, we have problem with these:

1) Czech have differences between genus in example "was" -> he - "byl" | she - "byla"
2) our pronouns have cases which are changing even with gender:

            I   |  you | He  |  She |  It
1. case   já   ty   on   ona   ono
2. case   mě, mne   tebe, tě   jeho/něho, ho   jí/ní   jeho/něho, ho
jej/něj      jej/něj
3. case   mi, mně   tobě, ti   jemu/němu, mu   jí/ní   jemu/němu, mu
4. case   mě, mne   tebe, tě   jeho/něho, ho   ji/ni   je/ně, ho
jej/něj      jej/něj
5. case   -   ty   -   -   -
6. case   o mně   o tobě   o něm   o ní   o něm
7. case   mnou   tebou   jím/ním   jí/ní   jím/ním

3) same to adjectives have cases for ex "concetrated" -> He - "soustředěný" | She - "soustředěná"

As rest of slovan language we can make it work but it wont be nice.

[ison]

I see. It's very tricky. But I'm not sure there's any easy solution we can use here. We could add a few extra custom strings, but there are thousands of verbs, so I don't think it would solve the problem. We'd need a new, special syntax in xml like:
<Message>John {PAWN_gender=="male" ? "maleVerb" : "femaleVerb"} something.</Message>

[davefin]

Tbh until there's a more robust translation system (and _if_ that is something Ludeon wants to dedicate resources to), I'd just accept the Czech and similar translations to be kinda inaccurate.

The problem is, it's not just the verbs that change form depending on context like the subject or object of the sentence - it's also adverbs, adjectives, and a bunch of others. English is refreshingly (and stupidly) simple in this regard.

E.g. if you want to say that 'he punched the green dog', you'd say 'on udeřil zeleného psa', while 'she punched the green cat' would be 'ona udeřila zelenou kočku.'
The main difference there is that the actor and subject are masculine in the first example, feminine in the second, no other changes.

It's kinda hard to illustrate without writing up a page of examples and rules, but basically, almost every word in a typical sentence changes form, typically in the prefix and suffix, based on the context - the gender of the subject, the actor, the tense, etc. We also have a neutral gender just so stuff isn't that easy lol.

(To answer Tynan's question specifically, the gender usually comes from convention - some make sense (a girl is feminine, a boy is masculine), but a lot of things are just kinda arbitrary, sorta like in German - e.g. a pencil is feminine (ta tužka), a pen is neutral (to pero), a knife is masculine (ten nůž).


This stuff could be manageable in alerts, messages and the like as the volume is kinda low and the translators can generally write up a sentence in full if they know the context (incl. the grammatical gender of the actor etc), but especially stuff like the generative text like in art descriptions and combat descriptions would be significantly harder imo.

As a band-aid fix, adding an 'a' to the end of a verb usually makes it feminine, but yeah, that isn't really a robust approach. It'd probably need a more complex translation system, inlining it into XML would probably end up a nightmare of spaghetti. A lot of services I've seen which don't have better translations yet use the form of <verb>(a), as in 'User A poked User B' -> 'Uživatel A šťouchl(a) do Uživatele B', implying it could be either masculine or feminine, but that would probably feel out of place in RW.

Welp, this turned out to be a wall of text, but hopefully it clears something up.


EDIT: I now realized that there's a new translation tool up, don't know if that addresses this or not - this thread seems to be pretty old.