Fellow Interloopers !
Does anyone have access to the game files and could extract the words of Atlas ?
I have of course wikipage (2019 edit) https://nomanssky.gamepedia.com/Atlas_(language)
But doing my 2 iteration voyage I found that Atlas path quests line was completely changed, and even words are completely new.
Fellow Interloopers !
I had been working on a spreadsheet in Excel, but made a quick copy of all the alien language to Google Sheets. Not as organised as my work in excel, but hopefully still useful.
NMS version 2,32, so reasonably new
Hah, thanks! I knew Devlin would be the one to respond here.
Is there a column missing though? (I‘m on mobile) I only see Constants/Enums and English strings.
Are languages just 1:1 word mappings? And related words (to dance, a dance, dancer, dances, danced, dancing, etc.) will end up totally unrelated alien words? I had really hoped they would have chosen a smarter implementation…
Basically, this is impossible to localise… The original English text is already written in a foreign sounding „telegraph“ style, but for players who don‘t set this game to English, 80% of the dialog must be gibberish…
Imagine you are a translator and you are told to translate free floating words like „content“ (satisfied or capacity?), „love“ (to love? The love? beloved?), „to“ or „at“ or „for“ (What if the target language has endings instead of prepositions?), „The“/„a“ (even some European languages don’t have any articles), „is“ (gender? number? tense? Auxiliary or main?), „has“ (own? Must? Perfective?), and what is one word in one language could correspond to several in another… You get the idea.
So HG thinks their dialog says (made-up example) “the content man at arms hit the bull’s eye” and it’s pieced together into other languages as “this capacity human near arms the impact this of cattle eye”…
I assume you think the alien language is missing, but this is actually it, just the ID and the English translation for the words that can be learned in-game. For most of the alien language, you will no longer find direct translations. The Atlas used to have full translated sentences in the language files, but even these have now mostly become part of the learning process.
The language is procedural with quite many steps and hoops, going through long strings to finally be reduced to limited alien strings that appear to be alien words. Nobody has been able to fully understand the process behind the language creation (yet). It is all done outside of reach of those willing to create mods. The proc gen for the language is all done inside the executable. Only one person has attempted to deconstruct the inner workings and has actually gotten pretty close. Maybe one they we can have full code where this full process is being replicated.
Does the alien language make any sense? Nope, not at all! The only thing that has been focused on is being able to generate a vast amount of words that look alien. There are some rules to make these words not totally impossible, but that’s about it. Variations of words that are similar or only grammatically slightly different, will end up completely different words. Even just changing capitalisation will have totally unrelated results. So yes, basically just 1:1 mappings to the input given.
To be honest, I believe the same procedure is used for all the known alien languages. I truly doubt they used totally different generation code for each. Possibly some initial alterations to have the distinction between race, but the end result is likely spat out by most of the same code. This somewhat works if you do not look into it too deeply, but in the end it is pretty much all the same. All that really matters, is the result looking alien to us, while somewhat being able to pronounce it.
Now to solve the actual translations, players have been comparing dialogue results, with partly translated results, to then being able to fill in the gaps eventually. Do a run without knowing any of the alien languages, make screenshots of everything, to then learn a few words, capturing once more to compare. Count of words in sentences has become important, because this never changes between translations. Eventually you can fill in all the blanks.
I have been gathering translated results over time from various sources. I have started attempts to map it all into excel sheets. The continuous changes doesn’t make this an easy process though. Not to mention the amount of time and effort required. I have never fully completed the dictionaries, but I am at a point where I can load the localisation files into NP++, do some quick regex procedures on them and import them into excel to generate linked libraries. All I have left to do is link up translations players have been making, to see how much is still missing. I bet you there is a lot missing, as over time each language has grown considerably. Still you only require 250 words for each species, minus the Atlas (which has a somewhat new implementation).
Edit: I forgot to mention that more data has actually been added to the dictionary, with fields such as ‘frequency’, ‘group’, ‘category’, ‘interact effect’, and ‘level’. I did leave those out in the Google sheet.
For those interested in Procedural Language and Interactive Storytelling, I found Emily Short having written quite a few interesting articles.
Thank you, very interesting, I didn’t know that. When the game came out there was a list on the wiki of alien words mapped to English words. And I had assumed that list was simply copied from the source code.
I understand that the words themselves can be procedurally generated. I am more interested in the sentence level (which was annoying/incomprehensible with localisation, at least when I last checked). I’ll switch my language settings and have another look.
Thank you also for your rich description of processes. But I still don’t understand something.
If there is a new story line from Atlas. And you can only interact with it once. So how to read what Atlas says to us at all ?
I haven’t looked into it enough to know if and what has changed. I do however believe we can still learn Atlas words, while waiting to follow the Atlas path. This would then give us a better chance of seeing words translated. It will be a tedious process, as it has been for most of the alien language in the past, and can most certainly still be done. Of course using some tools like a Save Editor and a combined community effort, will make this task a lot easier.