Review by asking, not asserting

I spent time in someone else’s design proposal this week. Nearly all of my comments were questions about boundaries: why the scope stops where it does, and which responsibilities belong in the new layer versus the one already behind it, given that some users will only ever touch one of them. One comment was that I couldn’t tell what a section was trying to say. Questions do more work than my opinions. They surface the assumption the author hasn’t tested yet, and they leave the design where it belongs. That reflex paid off again later in the week. A team reported an integration was broken in test. Before anyone escalated, someone asked when they’d last run it. They’d tested before the fix shipped. Nothing was broken. Saying “I can’t tell what this means” is not a weak review comment. It’s usually the most useful one in the document.