2026-06-14 To Type a Tale devlog
Addressing feedback
published: Sun 14 June 2026This week involved a lot of user-facing changes. That's what happens when several people try my game and give feedback (hint, hint)! The big-ticket items are the mostly-finalized graphics, new execute modes, and some convenience features. The big non-user-facing change this week is that I finally started formalized QA.
Let's start with those graphics: All of the main text is now displayed on a papyrus scroll which scrolls to display new text within the chapter, opens and closes to change chapters, and gets moved onscreen/offscreen to change stories (mostly seen when moving to the menu). I'm not entirely happy with how fast each of these animations are, but they'll do for now. The pager and banks are placed on top of what I internally call "agent paper," a sort of in-universe filing method, which means the player is actually a clerk typing up an official report from Tola.
The big new accessibility options are execute modes and rainbow modes. Requested by Matthew, execute modes allow the player to choose when blue words are executed (i.e. when the game processes them in the backend), when paging occurs after execution, and, when typing multiple blue words before execution, which blue word is chosen to be executed. As part of this, I also put a new section below the pager which indicates which blue word was executed/will be executed. The rainbow modes allow enabling/disabling the rainbow highlighting on typing, typed, flying, and jumbled letters individually. Try experimenting with only turning each one on or off at a time! You'll also notice that the flying/jumbled letters are at a lower opacity and placed underneath the other text, for readability (thanks to Matthew for the complaint/suggestion).
Some other important convenience features made it in while I was tinkering with shaders and code organization. Full words are now highlighted/collected when clicking, rather than individual letters, making clicking a viable alternative to typing (thanks to Ryan for the suggestion). A loading overlay hides the initial UI setup, because Godot has issues with shuffling things around after they're visible. The finished-story plaintext is now written out only when reaching an end in the story, but the first-restart message now mentions this and provides a link to the user folder. I'll probably provide links to said user folders in the menu eventually. Music now remains loaded (and thus, playing) in the menu until a new music group is loaded (thanks to Matthew for the suggestion).
Finally for the user-facing changes, a credits page was added! Several of you claimed you didn't do anything worth crediting, but as the above demonstrates, you very much did! Keep the feedback coming!
The major backend work I want to share from this week was putting together a formal QA process. I now have a Google Sheets document listing a dozen test scenarios with detailed steps, expected results, and last execution times/successes. The first time I ran it was yesterday (6/13) and it found several bugs, so it's already proven its worth. I plan to run the full suite (takes around 1-2 hours) every Saturday, or before public releases. As more feedback is provided, I'll put together new test scenarios to make sure future work doesn't cause regressions.
Some scripting work is also worth mentioning: I finally fully implemented the !pagebreak command, which forces the paging in the main screen to actually pause when gathering text. You won't notice it anywhere yet, as I'm not using it until chapter 2, but it also helps with the options pages. To help with formatting the credits page, I also tweaked the paragraph-processing logic to respect leading spaces, allowing those indents. Scripts can also now fire notifications which appear at the bottom of the screen. The only script using this right now is the tutorial (to provide a reminder about paging), but I will eventually use it for achievement notifications. I still need to figure out some pizazz to make them more noticeable.
I have also been working on some new fuzzy-search algorithms to allow some typing strategies you've asked for (e.g. skipping already-typed words), but I've had little luck so far. One of them (I've dubbed it "confidence searching") has promise, but also tends to hallucinate and fail to recognize letters you've typed. I have an idea for another (dubbed "heat-map searching"), but if that doesn't work out, I may just have to keep the old method and deal with the complaints.
Next week will be more one-off work, finishing addressing feedback comments and tying up loose ends before the public demo. The last major art work I have is the music for chapter 1, which has been looming over me since the start of this project. Fingers crossed!