Finish the underside — Steady<br>The clearest test case for a well-made product is the part no one will see.

A carpenter making a table will, at some point, have to decide how to finish the underside. The underside is the part that no diner will ever look at. The varnish you put there will not make the table look any better when it’s standing. The sanding you do will not be admired. Many respectable table makers even argue it will not meaningfully prevent moisture or improve the table’s lifetime. The table will look just as good on the showroom floor. The financial return on the underside-work is, by any short-term calculation, zero.

But you have to finish the underside.

When it comes to software, there is often this argument: this part won’t be seen and the software should run without it, so it doesn’t need to be done. It is a damaging mentality. The part that won’t be seen is part of the same thing as the part that will be seen, and to do one well and the other carelessly is to do neither well. The thing is whole. All of it has to be made with the same attention.

Modern instrumental thinking treats objects as bundles of features, each of which can be optimized separately according to whether it appears in the user’s experience. If a feature does not appear, it is wasted effort. If it doesn’t appear enough, it is acceptable for it to be rough, unfinished, less carefully made than the parts that do. This is the logic of the inferior table. It is also the logic of much modern making, especially modern software making, where what is not visible to the user is, by this calculation, not worth doing. The result is artifacts that work as long as you stay on the optimized surface and that crumble whenever you step off it. But when the user does something unexpected, the unmade parts reveal themselves, and the user experiences what was always true: that thing was never whole.

If this resonates, I write occasional notes on ADHD and building Steady. Sent when there's something worth saying.

Subscribe

I try my best to resist this. I tried my best to finish Steady’s underside not because the user will notice but because making a table is making a table. The completeness is not for the user’s eyes; it is for the maker’s self-image. I hope I did. Any corner that fell short of that would stay with me.

This also is a position about what kind of relation the maker has to the user. The maker who optimizes only the visible parts is in an adversarial or at best transactional relation. The maker who finishes the underside is in a different relation: I will give you the thing, all of it, including the parts you cannot inspect.

Users come to recognize, after enough exposure, which makers they can trust. They cannot always say why. But they know. The maker has been answering, in the unseen parts, a question the user did not consciously ask, and the user has felt the answer, and the respect.

Ahmed Soliman