What are the steps to follow for new translation of the game?

[absay]

Hi.

I'd like to start a new translation of the game for Mexican Spanish (es_MX) based on the already existing es_ES translation. In fact, I've been working on it for a while and I already sent a mail to the translator manager almost a week ago to get access to the Web Translator (I already have an Openttd.org account) but hadn't had any answer so far, so I'm starting to wonder if adding a new language would actually be possible so I can continue with the translation work, or otherwise stop.

In case you ask, yes, both dialects change in several ways, I'd say much more than en_US vs en_GB just to mention an example, so I think a whole new language is justified at least from the perspective of language differences.

I'm posting this here to know exactly what I should do now or hopefully the guy in charge of managing the translations will see it.

Thanks.

[Redirect Left]

I think all the information you need is here: https://translator.openttd.org/
I believe OpenTTD do all of its string translations through a web interface.

[absay]

Thanks for replying.

I have read "all the information" I need but my point is that this is for a new translation and the manager hasn't answered my mail to grant me access to the Web Translator (as I explicitly mentioned above if you'd like to re-read my post) to undertake the string translation.

So I wonder if there's another way of getting access to said website and adding a new language or if I'll have to wait until the guy responds.

[Redirect Left]

There are numerous developers for OpenTTD who frequently read these forums. There is a good chance at least one of them will see it in a fairly short period of time (I'd say less than a day). However I do not know which ones specifically handle translations, i'm sure they'll pass on the message its a fairly unified team with a lot of communication from what I have seen here on the forums.

My best advice is simply to just sit back, play a game of OpenTTD with your favorite drink and i'd be surprised if it took more than a day or so for a more helpful reply than I can give. I'm just a player unfortunately!

[absay]

Alright, I supposed this ought to be the case.

I guess in the meantime I'll keep externally translating stuff then.

Thanks for the answer!

[Supercheese]

You can also poke them [url=irc://irc.oftc.net/openttd]on IRC[/url], you might find that a bit more responsive.

[FLHerne]

Further to that, if you don't have an IRC client, click 'Chat' at the top of this forum and select 'OpenTTD channel'. It's the same place.

[stravagante]

When I applied to be a translator, I waited, I think 2 months or so for a reply. So, like someone said, relax and enjoy the game.
Also, I'm not sure that external translation will help you much, because web translator gives you random strings to translate. I mean you could always make excel "table" with alphabetical order of your translation, and then simply paste it where needed.

[absay]

Thanks for the chat suggestion. Unfortunately, I can't seem to catch anyone who is willing to help me because of the timezone differences. I guess I have to try earlier.

And I've been enjoying the game for about 6 years, so that's no an issue.

Also, I'm not sure that external translation will help you much, because web translator gives you random strings to translate. I mean you could always make excel "table" with alphabetical order of your translation, and then simply paste it where needed.

Yeah, I'm doing something like that though not precisely with tables. I'm not new to translation tools so I'm pretty aware of the random nature of the process. The point of an external translation is to have an organized guideline document which is as consistent as possible, something the existing es_ES translation lacks of to a certain degree.

[planetmaker]

Hello,

Hi.

I'd like to start a new translation of the game for Mexican Spanish (es_MX) based on the already existing es_ES translation. In fact, I've been working on it for a while and I already sent a mail to the translator manager almost a week ago to get access to the Web Translator (I already have an Openttd.org account) but hadn't had any answer so far, so I'm starting to wonder if adding a new language would actually be possible so I can continue with the translation work, or otherwise stop.

In case you ask, yes, both dialects change in several ways, I'd say much more than en_US vs en_GB just to mention an example, so I think a whole new language is justified at least from the perspective of language differences.

I'm posting this here to know exactly what I should do now or hopefully the guy in charge of managing the translations will see it.

Thanks.

Sorry, that would be me who had been a bit negligent in that respect in answering translation e-mails fast.

The general procedure in adding a new translation (and that's what es-MX would be) is as such: we ask the translators to provide a rather complete translation into the new language and send us that language file (via our bug tracker, or to the translator e-mail address). We will then commit that initial translation and assign the translator to that so that s/he can maintain it for the future.
As starting point I suggest you take the es_ES file and adopt it: http://hg.openttd.org/trunk.hg/raw-file ... panish.txt
Not sure whether the number separators are treated differently in Mexico than in Spain, change that appropriately, if necessary.
We'll look up the winlangid and decide on the grflangid. Plurals and genders certainly are the same as Spanish (ES).

So in essence: yes, please continue your localisation effort in creating the es_MX translation, we can add that to the game in due time.

Cheers,
pm

[absay]

Looks like I'm on the right track (I'm sure this pun is already boring here?).

Regarding the "negligence" you mention, not at all. I started to feel a bit pushy when people told me a response could take months, so one week didn't seem to be an appropriate time anyway.

And yes, digit grouping does change, in Mexico we use the same as in UK or US; genders, like you said, are identical, or at least I can't come up with any exception.

Thanks Planetmaker, I will continue with the translation, the es_ES document you liked is the one I've been following, along with en_GB to retain as much fidelity to the original as possible. Cheers.

[absay]

Alright, it took me a little longer than expected, mostly because it was thoroughly proofread, but it's finished. I already uploaded it to the bug tracker (not sure if the "bug" details selection was correct though...). Anyhow, I attached a copy to this post too.
Looking forward to knowing the next step!


As a side request, I'd like to translate FIRS now but I imagine I have to wait until the actual translation of the game gets added to the trunk, right?

[frosch]

Thanks.

I'll try to add it over the weekend.
First OTTD trunk, then Eints, then DevZone, then various NewGRF Wikis and Tools.

For FIRS translation, you can already register to DevZone, if you haven't yet.
We'll then add you to the translators groups when it exists.

[GunChleoc]

I really don't understand why people have to deliver a 100% translation before they get access to the translation tool. I was kindly let in after a partial translation, as I was ready to pull out my hair... I translate per terminology item to keep everything consistent, and not from top to bottom. There is no way to track your progress in a text file.

I understand that the devs don't want to waste their time with dead translation projects though.

[absay]

I understand that the devs don't want to waste their time with dead translation projects though.

Agreed, but 100% is a little too much. I mean, the translator is supposedly to finish it either way but 50% or 75% should do to prove the translator is serious about the project.

And speaking of the translation tool, the only thing I dislike about the web translator is that it will give you random strings to translate, I would prefer to go through the file in order but I guess this cannot be changed.

Also, I finished translating FIRS and tried to start with ISR but I couldn't create the project for my language, so I went to the devzone and I noticed I'm not a member of the translators group, so I opened a new "membership issue" and I'm waiting for approval or something.

The question is: will I need to do the same for every project I'm not part of?

Another question: what about the projects that are not listed in the dev zone, namely Japanese Tracks, for example? How do I go about those? Do I need to to contact the author?

[GunChleoc]

And speaking of the translation tool, the only thing I dislike about the web translator is that it will give you random strings to translate, I would prefer to go through the file in order but I guess this cannot be changed.

+1 - the random order is robbing the strings of their context.

How easy this is to change depends on whether this was done to make implementation easier, or because somebody thought it was a good idea at the time.

[Alberth]

And speaking of the translation tool, the only thing I dislike about the web translator is that it will give you random strings to translate, I would prefer to go through the file in order but I guess this cannot be changed.

Everything can be changed, but it's work, potentially a lot of work.

The tool has no concept "next string". Each request is like the first one, it doesn't track what you're doing at all, it doesn't know if your entry is the first or the 73rd submission. Any solution that relates the next query with the previous one will thus not work. Similarly, the tool has no concept of different translators, all translators are equal. If multiple people translate in the same project for the same language, the tool does not differentiate between those users, each gets a new string when he/she asks for it, without knowing or remembering what was given out to other people.

There are 2 problems going down sequentially through the file. The first one is collisions. If two people translate in the same project for the same language, and they submit a translation for the same string, the first submission will win and the second submission is rejected. When going through the file from top to bottom, the hit rate of this problem is 100%. The tools gives out the same string to everyone until one gives a translation, then it gives out the next string to everyone, until it gets a translation, and so on. With random strings, this problem practically never happens.

A second problem is that you cannot skip a string anymore. The first untranslated string stays the first untranslated string, even if you don't want to translate it now. Effectively, you cannot proceed at all at that point until you do that translation. With random selection, you'll hit the same string much less often, and you can continue translating other strings.

Finally, authors often put a lot of very similar strings after each other. It's easy to make an error in these subtle different strings. By randomizing you get a more diverse palette of text strings, increasing the odds of catching copy/paste errors.

The question is: will I need to do the same for every project I'm not part of?

Depends on the project. There is a global translator setup. If project attach to that, all translators of all languages immediately get access to the project. Some projects however either attached to the setup in the wrong way, or they want to have more control over who is translating for them.

Another question: what about the projects that are not listed in the dev zone, namely Japanese Tracks, for example? How do I go about those? Do I need to to contact the author?

Likely, but there are 2 forms of "not listed". If you mean not listed in the translatable projects, but the project is hosted at the devzone (see the "projects" list), then the project is not attached to the translator setup. Reasons can be anything from not using the strings as done by the translator tool (eg a different format, NFO does things differently, afaik), to not knowing it exists.

If the project itself is also not at the devzone, the authors preferred to use something else to host their code and strings.

[GunChleoc]

Obviously a lot of work then to un-randomize the strings.

There are 2 problems going down sequentially through the file. The first one is collisions.

Which could be solved by assigning any string that a translator has fetched to that author with a timeout attached. Yep, that's quite a bit of work to program as well.

A second problem is that you cannot skip a string anymore.

Not quite true. You can select any specific string from the string overview right now, e.g. to fix translation errors, so this is mainly a matter of the user selecting their first string to start translating and then navigating the strings from there.

Finally, authors often put a lot of very similar strings after each other. It's easy to make an error in these subtle different strings. By randomizing you get a more diverse palette of text strings, increasing the odds of catching copy/paste errors.

Yes, these do happen - I wonder though if they are frequent enough to cause more translation problems than lack of context does.

[Alberth]

Just a few notes to ensure we are at the same page.

There are 2 problems going down sequentially through the file. The first one is collisions.

Which could be solved by assigning any string that a translator has fetched to that author with a timeout attached. Yep, that's quite a bit of work to program as well.

Translator queries are currently not tracked, the tool doesn't know what your last string was, or your last 5 or so. Having no state in the translator tool was one of the first decisions. It simplifies the design and makes that you can pull down the translator tool at any time.

After solving that, you run into edge cases where there are less strings than translators, or the same translator doing more queries.

A second problem is that you cannot skip a string anymore.

Not quite true. You can select any specific string from the string overview right now, e.g. to fix translation errors, so this is mainly a matter of the user selecting their first string to start translating and then navigating the strings from there.

This assumes tracking of what string was given out to who, when, which isn't done yet, see above.

If you want to have a go at it, feel free, the source is available at the devzone.

[GunChleoc]

Yes, we are on the same page that it would be a lot of work. So, we won't have this any time soon, unless somebody has enough interest to spend a loooot of time on it. I'm already snowed under with projects, so no can do