Name things the way people use them

I spent a session in Olympus, the admin platform behind Frontlyne, rebuilding an image feature I had shipped the day before. The rebuild was smaller than the original, and most of the good decisions were about naming and honesty rather than code.

Name by shape, not by platform

My first version of the format picker was organised by where the image would be posted, with options like “Instagram Post”. I changed it to name each option by its shape instead: “Square”, “Portrait”, “Story”. The reason is simple. The same image ends up on lots of different platforms, so the platform is the wrong label. People are choosing a shape, and the picker should match the way they think, not the way the file is eventually used.

I went one step further and made each option a little rectangle in the right proportions, so you pick the shape that looks right rather than reading a number and doing the maths in your head. It was a tiny change that made the whole thing easier to use.

A fake stand-in is not a shortcut

The engine that actually makes the images was not built yet. The tempting move was to wire up a fake version that returns plausible-looking results so the screen “works” end to end. I chose instead to show a clear “not connected yet” message.

A fake stand-in does not age well. It survives longer than it should, it hides the bugs that only show up when the real thing is connected, and it gives you the comfortable illusion that something has been tested when it has not. A loud, honest “this is not wired up yet” is better, because it forces the real connection to happen before anyone can claim the feature works.

Both decisions come from the same place: build things that tell the truth about themselves, to the people using them and to the people building them.