Journal 📝

Sunday was all sound, and the sound is where the game actually happens.

Saturday, first day with actual code. The pitch from Friday fitted in one line: you are a rock, you get wings, and the wings only last while you keep beating them in time. One button and nothing else.

Friday was a chill day. No code, just sitting with a theme I did not understand :)

Good news: we published in time. Bad news: Unity takes HOURS TO COMPILE D:

First wasm bundle I built was a clean 370 MB. A piano sample is a 10 s stereo 44.1 kHz WAV and we ship one per key. Not quite portfolio material :upsidedown:

The old Unity Ivy is still up on the main page. Underneath I’ve been rebuilding it on top of Simurg, a custom engine I’m writing in C :)

I assumed wasm on mobile would Just Work. Yep, very naive of me.

If I’m going to spend a year writing this thing, I want the iteration loop to feel instant. So every layer of the engine hot-reloads on save:

Until today a dimension was a single key, so something like Spain on version 2.1 was impossible without creating a new materialized view for that exact combination. Dimensions can hold several keys now, and the event type is always prepended, so you can still slice everything by type. Stats queries got GroupBy too: filter by one dimension, aggregate by another.

Events are clickable everywhere now - tables, lists, timelines - and they all open the same detail drawer, with its meta, the session timeline and the raw data in scrollable sections.

There is one Docker file for the whole server now, with ARG APP_NAME picking which binary to build and BuildKit cache mounts so it doesn’t compile everything from scratch every time. Sixteen almost-identical Dockerfiles is sixteen places to forget a change.

Small day. Event summaries can be filtered by country, and by more than one at a time.

Event summaries: endpoint, migration, store and UI. One row per event type with its totals, instead of asking for the raw events every time you open the page.

IconSelect is now an emoji picker with search, categories and a virtual grid. The virtual grid is not optional, rendering every emoji at once makes the whole page crawl.

Moved the secrets and the config to shared Terraform resources. Every chart had its own copy of the same values, which was fine with three services and stupid with sixteen. The Helm charts got a lot smaller after that.

The journeys page is live. It draws a Sankey chart with the paths people take between events, so instead of counting how many times something happened you can see what happened before and after it.

Last day of development: time to finish everything and hope it compiles (spoiler: it didn’t).

Main day for development. Everything was decided so big coffee and let’s go!

First day of the jam, mostly to discuss what kind of game we want to build. Spoiler: cozy game!

After dedicating time and effort to enhancing the user experience, I’m happy to share the details of the latest update for my I-Ching app.

After months of juggling my day job, life’s curveballs, and countless cups of coffee, I’m thrilled to announce that the new version of my I-Ching Oracle app is officially live!

We didn’t actually have the time to discuss some conclusions, or have some kind of post-mortem… but more or less we agreed on some conclusions.

This is the final day, this is the time to rush D:

I slept quite much for jam-standards (6h or so), since I’m a bit old for non-sleep jams :D.

This was the first day! Lots of ideas, and nerves. It was not our first jam game (I even did other jams with my brother Ale), but it was the first one with my cousin Miguel. The first one with a proper designer!.

I lost the source code of the game! D: - what happened next will surprise you!

I’ve changed the font of the website, and I think this one looks way better than the old one.

I’m moving from Nuxt.js (Vue) to Astro.

I’ve just discovered that lazy loading components is not that great for this kind of apps.

I’ve started the preparations for the release in android apps. It was not straight forward, and I had to dig a bit into github issues, but I think it’s now ready to release :)

I’ve finished all the preparations to finally have a first release. For now it’s just a Web release (as web app and PWA). I had to tweak a lot of stuff before making this production ready. I had to generate icons, manifest, PWA stuff…

I’ve created a new view to explain what the app is about.

I’ve finished the journal view (where you can check previous oracle results), and I’ve tweaked some views.

I’ve finished the settings page and added more (english) texts. This day was not that fun, since I had to add a lot of texts for the hexagrams (quotes, descriptions, etc.).

I’ve spent the day adding graphics (Hexagram) to the results.

I’ve finished the main app flow. It’s far from ready, but it’s something! \o/

I’ve been testing whether Ionic works as expected for all the platforms I’m targeting and… Ionic it is!

I’ve spent the whole day researching what’s the best technology to achieve this. This is meant to be a very simple app, so we don’t need the full power of Native apps, but… maybe it’s easy to build those for any platform too?