Tek'Notes !

You leave a big session in front of a magnificent wall of sound. One track floored you, and the next day you can’t find out who was playing. The name is stuck in a story that’s already expired, or in your phone notes between a shopping list and a Wi-Fi password.

Music apps aren’t much help here: a good part of this scene isn’t on any platform.

It’s the same idea as in Swikko: self-sufficiency (“autosuffisance” in French). Nobody fills the catalogue in advance. When someone logs a soundsystem or an artist that isn’t in the app yet, they create the entry, and it becomes available to everyone. Every entry can be corrected like a wiki, with its history¹².

There’s no administrator approving each addition. The app warns when an entry may already exist¹¹, the community can propose a merge, report¹³ an entry or block someone, and a heavily reported entry hides itself.

You start with a short introduction⁰¹, then the notebook⁰³ fills up night after night.

• log a party in two taps, or by voice, with the venue, the soundsystems and the artists;
• add to the night⁰⁵ a set (an artist on a soundsystem), a rating from 1 to 5 and a comment;
• keep everything offline: the notebook works without an account and without a network;
• find a name by place, artist or date, even with a typo⁰⁴;
• protect the app with a code⁰² on opening.

The public directory⁰⁷ is shared: you search it, and with an account you create and correct entries. Each entry⁰⁹ shows its aliases, styles, artists and the nights shared by others.

I chose Kotlin after a very interesting class given by a visiting lecturer who explained things really well. It gave me the idea of writing the core of the app once, in Kotlin Multiplatform: the models, the search, the rules and the local database.

On top of it, two native apps: SwiftUI on iPhone and Jetpack Compose on Android. Each screen is written once per platform, but the logic exists in a single copy.

On the server side, Supabase (Postgres) with access rules written in SQL and tested automatically on every change. Changes made offline wait in a local queue and go out as soon as there’s a network.

The first version is written and tested, and I’m trying it on my iPhone with demo data that is clearly fictional: it’s what the screenshots on this page were taken with. It isn’t published yet: I still have to set up the production server and get the legal texts reviewed. The code is public.