Primitive design and how I’m spending my week
20.06, Thursday 9 Dec 2021 Link to this post
I’m currently trying to wireframe a new service. Precisely what it is doesn’t matter right now, but it’s software.
The work is less about “this is the design” and more that I’m figuring out whether this is a viable starting point for the actual design work which comes later.
On my paper I have two lists. One is a list of top-level features. The second is a list of desired user outcomes (things like: “have a sense of familiar strangers”). And then I have four sketched boxes which represent screens, and boxes on them. It’s a pretty simple service at its core…
…and it’s built out of an existing platform. It’s an application of that platform. The top-level features already exist; hitting the specific outcomes will steer future development (and that’s part of the purpose of this new service: a concept car kinda).
Then I have additional requirements for the interaction design:
- I want it to feel intuitive
- I want any new features to be platform features, not one-offs.
And the second of those is weird, right? It’s like sketching out a toy spaceship, having a list of rules about play, and attempting to simultaneously invent the shape of the Lego brick.
That’s platform design I suppose. Redesigning a newspaper will means bouncing between comps and style guides, designing both. Inventing the iPhone user interface will have seen apps and app paradigm evolving together. Those are examples much bigger than what I’m attempting.
Actually what I’m reminded of is this: eye-balling a graph during my physics undergrad and figuring out how to express it in simple mathematical terms, then using that as a model to make predictions, and testing and refining that model. The process of expressing a system as composable primitives.
I’ve talked before about primitives and notation (August 2021) and this process is part of that same puzzle: I want my primitive building blocks to feel natural and well-rounded and logical; I want my service to make sense and achieve its aims and come together neatly from the primitives. There’s an iterative back and forth to get there, between these two orthogonal descriptions of the same thing.
It’s enjoyable work and that’s mainly how I’m spending this week.
I’m sure other people have described this kind of process. I’d like to read about it if it rings any bells for you.
Sorry to be so cryptic.