ITCH | MEGA | ITCH-DELUXE | MEGA-DELUXE

Later update than usual as I have been working till the last minute on getting our community translation support into a place where we can do some first experiments on. But one thing after the other.

First of all, there is no new content this week as we have been working exclusivively on technical items in order to improve our technical infrastructure in preparation for setting up the prototype for Liora. I wanted to release a 1.1 version of our RPGMaker framework AuraMZ today as well, but with all the technical changes, we still need a bit more testing and cleaning up to get there.

Trihan has been busy reworking out quest system to make it more robust and friendly for translation - those who have an interest for the technical details will find a new Quests.json in the data folder of the game.

Besides refactoring work, as mentioned above, we were also working on pushing forward the translation support. In theory we have reached a point where all strings are now localizeable. Before, we wore mostly working on making it so our plugins would no longer use hard coded strings but instead get them through a localization engine. We also worked on automatically detecting all strings that need to be exported out of the game into translation files. This week we implemented the other direction, namely for translated strings to be properly loaded and displayed in-game. If you make a change in the languages/de you will see the change in the game now.

Moreover, I have also begun to setup the infrastructure for collaborating on community translations. In general, we have two directions here.

(1) Work directly on the Translation GitLab https://gitgud.io/auragamedev/star_knightess_aura_translation

I have created a new git project which contains the extracted translation data. Translators would be able to use a translation tool of their choice (for example POEdit, but in practice the choice of the translation tool is up to the translator) to contribute with translations. Since the infrastructure is originally meant for developing code it would also be possible to do reviews through Merge Requests.

The downside of this approach is that there is little high-level translation support out of the box and we would need to make additional functionality ourselves. I have quickly hacked together a simple script which gets executed whenever a change is merged into the main repository and builds a very ruidementary dashboard

https://star-knightess-aura-translation-eaca84.gitgud.site/

so interested translators could get an overview which files are still in need for translation. (Note that the German percentage here is incorrect since it identifies untranslated strings as them being empty but we initialized the German language by copy&pasting the base english one. In contrast to the French files which I generated through a script that creates the data in a clean state).

Another downside is that the tooling requires the user to deal with "software developer tools" like git. On the other hand, it would be very easy for non-translator developers who want to supply scripts to help out to do so.

(2) The second option is to use a cloud solution. The first big hurdle here however is that SKA is a HUGE game. 900k words. I checked out the plans for most crowd translation services and at first it looked promising but then I found out that their "costs per word" didn't refer to the original word count but to original word count x number of languages. ...Which quickly escalated the prices. Those services are definitely not meant for story heavy games.

Moving on, the next path would lead me to the option of self-hosting such a service and that's what I did. Among the possible candidates there was only one who was really viable - weblate.

https://translate.starknightessaura.com/projects/star-knightess-aura/

The current instance doesn't feedback changes into the main repo (but is connected to it) so you can play around with it. However, apparantly adding languages doesn't work and I spent most of today despairing over being unable to get emails working (which are necessary in order to send account activation mails).

This has very quickly revealed the major downside of this approach: It's a huge software that's annoying to maintain for a single person. Especially since I have only very little knowledge of hosting. And for some reason a lot of the things aren't configurable, those trying to use the registration function may notice that it asks for the person's "Full Name" - and I can't turn that off for some reason...

The software itself also kind of feels overwhelming and complicated to me, and while it's theoretically open source, in contrast to writing small python scripts, customizing this would definitely be out of scope. On the other other side, it's a tool with all the translation features one could wish for - tracking, plugins for machine translation suggestions, possibility for reviewers to review translations at any time, fields for discussion, progress bars everywhere and and and.

I have also opened up a dedicated channel for discussing this on Discord. Also is it just me or is it kind of weird that Patreon made it so that links become tiny. But ok.

Anyways, that's it for the the week. I hope the refactoring work won't cause too many bugs and until next week I will be starting up the development of the Liora prototype. There will be a devlog post to report on my progress but for actual builds, we will probably only do a prerelease build at the end of April since it doesn't really make sense to build one when I'm in the process of setting up a new game.

CHANGELOG 1.1.0-prerelease-5 (28.03.2025)

• Refactored Remove Collar action in auramz obsatcles plugin into a parameter
• Implemented JSON based loading of translation data
• Filtered out item obstacle choices if number of required items exceeds 99
• Disabled obstacle choice skills if skill is sealed
• Implemented compatibility functionality in mod_loader plugin for translations
• Refactored quests plugin to use external data file for structure instead of creating data at runtime
• Applied advanced enemies pattern note tag
• Integrated improvements made on AuraMZ into SKA
• Simplified more CTE code
• Extracted language codes from plugin into languages folder
• Added CTE parameter for configuring custom rules
• Implemented CTE parameter for extra databases and register compendium in it
• Implemented language font size offset field
• Extracted quests into external data, prepare them for CTE translation and update the quests in events
• Modied quest strings to use CTE localisations
• Updated quest events to use new structure

Balancing

• None

Bugfixes

• Added missing maid fulltimejob flags at Arwin's
• Fixed compendium data no longer matching schema
• Fixed repeated backlog entries
• Fixed incorrect export of escape commands in item and skill names
• Fixed CTE default font preloading not working
• Fixed remaining _status objective checks
• Fixed incorrect Colored Names parameter name