From the early prototypes {problems: and causes/resolutions}

• Batteries popped out of their brackets in the laser module: needs better mechanical design for use cases where it gets banged around in a bag with other sports gear
• Buttons triggered randomly even with internal pull-ups specified: some pins on the ESP32 are read only and have no pullups.
• "armed" LED didn't light up: bad solder joint with extremely cheap dupont wires never really wicking solder properly
• ESP32 won't come up in AP mode (crashes and restarts repeatedly): some clues here, but not conclusive. Moved some code out of the init_AP method, and it works, but I don't know why.
• Text doesn't fit on the display: probably should have done a separate prototype here to quickly work out the sizing
• The timing gate wouldn't start without a network connection: it worked fine at home... created a network hotspot on my phone with same config to work around.
• Debouncing the beam at the start to avoid stopping the timer prematurely: caught in early bench testing
• Beam state is read by polling: this is still a problem since it can occasionally miss an event (athlete finishes and only interrupts the beam for ~0.1s)
• No access through the housing to the program & reset buttons on the dev board: sidecutters
• Mounting the boards inside the housing: reprint & re-drill until the holes line up nicely (the "cut twice" approach)
• The housing for the laser is very hard to slide on/off: reduce the size of the bumps for the snap-in
• When a new client connects during a lap, their timer starts from zero: software bug, currently unresolved
• The mounting for the tripod shoe on the gate housing covers the holes for the ESP32 mounting, but has to be installed first: step 1, drill; step 2, better design
• The tripod shoe fits my tripod, but not others: the screw is a standard size (1/4"-20) but the shoes aren't; next design should use a knurled inset

There will be others.