Devblog #58
I bet you are a bit tired of constant "I did that" and "I am working on this" updates, so I thought that maybe I should switch it up a bit and talk about some general issues that might trouble the development of a game. Like for example, something that I never expected to be an issue at all - translation.
Right off the bat I knew I would make the game in English and Russian, those are two languages that I know well and I can be certain that I could make the translation myself. And most importantly, I thought that that should be enough. But as time went on, more and more people started asking for other languages. Spanish, French, Chinese, German, and I even had one Vietnamese request. Even though I did expect to translate the game to other languages eventually, I was not expecting people to actually ask for translations. But regardless of my expectations, I am now faced by a conundrum: try working on translations to other languages, or leave things as is and ignore the plights of potential fans.
One might think that there's no problem with translation nowadays with the advent of LLMs and them being able to translate large swaths of text with great effect. I would not lie, it would be nigh impossible for me to translate all the text in the game just by myself, so I am using LLMs to help with translation, but all the text that it outputs needs to be read through, checked, and more often than not - re-written. Sadly, you just cannot trust the machine to do a complete translation right now.
Some might say that that would be "better than nothing", but those people have not read as much of the LLM-translated text as I had. I still did consider that option, but then I read about Silksong's release being ruined for Chinese players because of not so nice translation, and I thought to myself "And that happened to a game that is not 90% text"! So that idea has been put on the shelf for now...
Another option, of course, is to find people to translate the game, or at least check the LLM's translations for me. But that poses another issue. A lot of new events are being written and revised by me up until the very day of the release. I could theoretically send something to a translator to finish within a day, but that would only be possible with paid translators, because I don't think it would be fair to impose this kind of schedule upon volunteers. And I definitely could not just release new events in some languages and not others, so every added language would make the number of dependencies grow.
Arrrrgh. Yeah. So this is what I am thinking about now. Maybe the solution will come to me next week.
I'll see you all then. Shady Character One, out!
Get Under Your Spell
Under Your Spell
Status | In development |
Author | Shady Characters |
Genre | Role Playing, Simulation, Visual Novel |
Tags | 2D, Adult, Dating Sim, Eroge, Erotic, Narrative, Romance, Sandbox |
Languages | English, Russian |
More posts
- Devblog #595 days ago
- Ilias: Legends of Alcyon by OzeCat6 days ago
- The Adventure Continues in Sekira Part 2! 🌟 by Mensh19 days ago
- Devblog #5720 days ago
- Version 0.3.3 Now with Journal!23 days ago
- Tap-Tap Legends!26 days ago
- Devblog #5627 days ago
- Devblog #5534 days ago
- Version 0.3.244 days ago
Comments
Log in with itch.io to leave a comment.
A terrible quandary.
If it were up to me, i wouldn't use AI at all. Work for humans should be made by humans, otherwise why even should art be made?
On the specific issue of translating, I'd go with volunteers. Even if the translation comes out a little later, I'd be happier with the better quality.
As a consumer, all I would want is a clear time frame: ie English comes out today, Italian & Chinese should be out by the end of the week. There's been a week delay on the French translation due to exams. Something like that anyway. Communication is key.
Good luck on your project, it seems to be going quite well.
This was very interesting to read, I never thought of such issues
Personally I'd use the LLM translations to start, then send them out to any volunteers to be edited after the two main translations are done and the next build is released. This takes the majority of the burden off your shoulders, and while it means the other translations see a slightly delayed release and a temporary reduction in quality, it also means you're not acting like a slave driver to any volunteers. Like any compromise it has its upsides and downsides, but to me at least it seems to be a sound idea.
Why not release languages after the game is finished?
Well, yeah, that's definitely always an option. But for now, more languages = bigger potential audience. Hence the dilemma.