Sh*t & Iterate
Ok, so let me just say it…
I’m no fan of watching users get dragged kicking and screaming through dozens of software updates while some software architect wannabe searches for success with the industry equivalent of a divining rod.
Can you reach a rough optima that way? Sure.
Successive approximation is an approach almost guaranteed to succeed. It’s seductive precisely because even the most inexperienced development team is almost certain to build something useable — eventually. Given enough time. And budget. And customers/developers willing to put up with all those dead ends, directional shifts, API changes, etc.
Whether “S&I” will work for your project depends on how much fire-retardant insulation you have between you and the user/developer you’re building for.
For example, Google could swap out pigeons for finches and who’d really notice? The user’s experience would be the same white page with a text field and a search button. On the other hand, every time Rails/Angular/React et. al. change direction they take an entire community of developers into the weeds with them. What’s the difference? The thickness of the insulating layer.
Google has an immensely thick insulating layer…no user has a clue how that stuff works and, more importantly, they don’t care. Enter text, click Search, get…
