This is what it looked like. A typical, boring DOS interface.
+ ++ Dr. Sbaitso would tell me what to do and I’d tell him about my 9-year-old problems. At the time, I was usually hungry or tired or bored. So I would answer in one of three ways. +
+ ++ And then he would ask, “Why are you bored?” or “Are you sure you are bored?” And then I would say, yes, I’m sure and start repeating myself. Next thing I know, he’s shouting a zen koan at me like this: +
++ (image: interfaces15.png alt: zen koan"> +
+ ++ And this would result in me having a tiny existential freakout. I would get sick of that topic and switch over to cussing or saying something inappropriate. And if I did that long enough, Dr. Sbaitso would throw a parity error. +
+ ++ The reason I bring this up is that it shapes how I think about interface writing. Dr. Sbaitso was pre-programmed with all of these different strings to respond with. +
+ + + ++ And while most of them were cold, poorly punctuated, and hilariously impersonal, the model of prompting the reader—this kind of call and response—really worked. It felt like a conversation. Maybe it was a little awkward, but it was a good start. +
+ ++ When you’re writing for the web, you’re having the same sort of conversation with your readers. You’re telling them to do something or asking them a question. Above all, you want it to make sense and feel natural to them. Our strings have to be useful—not funny—so we need to do the extra work of figuring out what our readers need. That makes it easier to show people around, ask for more of their time, or get them to take a particular action. +
+ ++ In his essay We Have Always Coded, Tim Maly says: +
+ ++ “It is no coincidence that many women have compared weaving code to instructing a child. With both kids and computers, you must carefully think through what you want them to do, and then carefully phrase your commands.” ++ +
From a high level, these are my goals when I'm writing strings: