Say you’re trying to test whether people like pizza. If you serve them burnt pizza, you’re not getting feedback on whether they like pizza. You only know that they don’t like burnt pizza. Similarly, when you’re only relying on the MVP, the fastest and cheapest functional prototype, you risk not actually testing your product, but rather a poor or flawed version of it.
Something strange is happening in the world of software: It’s slowly getting worse. Not all software, but a lot of it. It’s becoming more sluggish, less responsive, and subtly less reliable than it was a few years ago.
In some ways this is hyperbole. Objectively, we’ve never been able to do so much, so easily with our smartphones and laptops and tablets. We’ve never pushed more data between more places more readily. But while the insidious “worseness” I mention falls only in part on the engineering side of things, it falls harder on the more subjective, craft side of things, making it all the more worrisome.
Why should we care about this? Because the majority of our waking hours take place within the confines of applications. A truth recently amplified by the covid pandemic.
And I believe software used by millions (if not billions) has a moral duty to elevate the emotional and intellectual qualities of its users. That elevation begins with craft.
In the same way that physical architecture can affect a mind, so too can software. Slower, less reliable software is like Penn Station: Sure, you can catch a transfer from one train to another but the dreary lowness of the place, the lack of sunlight or sensible wayfinding will make you feel like a rat, truculent and worthless, and worse: You’ll acclimate to that feeling and accept it as a norm.
Hardware has literal and metaphorical edges — it must be fully complete and largely bug free to ship. Software? It’s far more amorphous, like mist. Patches can be endlessly pushed. It never ends. Faulty hardware can destroy a company. Faulty software can be patched. The butterfly keyboard debacle may never be lived down. Even as I type on this improved Magic Keyboard, I can’t help but wonder: Did they really test this thing? I had three butterfly keyboards die on me, twice in the field. Not fun. Hardware failures live long in the mind.
macOS software that adheres to craft — Things or Carbon Copy Cloner or BBEdit or Sublime Text (which, despite not being “native native” feels so solid and so responsive you’re willing to overlook its quirks) or Bear or Alfred or iA Writer or Keynote (arguably one of the best pieces of macOS software of all time) or anything by Panic, heck, even Terminal or Quicken (which, against all rational expectations is just a joy to use)5 — exists in troves, the existence of such proves to the Slacks or Twitters or Adobes of the world that it’s not impossible nor rare to produce craft-oriented software in service to user fluency, and still make a profit.
In fact, there’s a business case to be made for being craft- and fluency-focused. We’ve seen entire companies with business models that could be summarized as “Bloat-Free X” emerge in recent years. Affinity is bloat-free Adobe. Install Adobe Creative Cloud on your laptop and marvel at the no fewer than a dozen processes whirling around in the background for unknown purposes. It’s no surprise Affinity Photo and Publisher and Designer have taken off. Sketch’s main feature for many years was simply: Not Adobe.
And the web! When you care — when you really give a shit — the web is awe inspiring. I still can’t believe Figma is web-native (also born from the Not Adobe camp). That an application can feel so powerful, so fast, so well-crafted and be fully web-based should be a kind of lighthouse-archetype for all other sites lost in a sea of complexity and muck and unnecessary frameworks.