Craft2 min read
Software that outlasts its makers
Apps come and go. The things you put into them shouldn't. Notes on designing for the long run, from export formats to shutdown plans.
Here is an uncomfortable fact about software: most apps you use today will not exist in fifteen years. Companies are acquired, products are retired, formats fall out of support. Usually that's fine. But some things — a decade of journal entries, years of financial history — matter long after the app that held them is gone.
We think software should be designed with that ending in mind from the start.
Open formats, not exports-as-afterthought
An export button that produces an undocumented archive isn't really an exit. So every Selenra app treats export as a core feature, held to the same standard as everything else:
- Almanac exports entries as plain Markdown files, one per day, with photos in a folder alongside. Any text editor written in the last forty years can read them.
- Perigee exports your full history as CSV and JSON, with column names that make sense without a manual.
- Nocturne keeps its sleep log in a simple, documented format, because your notes are yours.
If you can't leave easily, you aren't really choosing to stay.
Offline-first where it matters
A meditation track that won't play without a network connection, or a journal that can't open on a plane, is a small failure of respect. Where an app's core job can be done on your device, we do it on your device. Sync is a convenience, not a dependency.
A plan for the ending
It's unusual to talk about shutting a product down before it launches. We think that's exactly when to do it — while there's no pressure to soften the promise.
So here is ours: if a Selenra product is ever retired, we'll give at least a year's notice, keep exports working throughout, and make sure everything you've created leaves with you in a readable form.
The long view
There's a phrase we keep coming back to in design reviews: will this still make sense in ten years? It's a useful filter. It argues against trendy interfaces and clever lock-in, and in favour of plain formats, quiet defaults, and features that earn their place.
We'd like to be around in ten years, too. But the more important goal is that your data would be fine either way.