when real users tried my thesis project

MySched started as my thesis. the first version was simple. scan a reg card. pull out the schedule. show it.

demos went fine. defense went fine.

then real students used it.

that was different. my phone, my lighting, my wifi were not what they had. old cameras. crumpled cards. bad angles. slow campus internet during enrollment.

stuff that worked in my tests broke on real photos. not because i was dumb. because i tested the easy cases.

when one person uses your app, weird bugs feel rare. when a hundred students use it the same week, weird bugs are the whole week.

so i stopped chasing shiny features for a bit. i cared more about things not breaking.

if a reminder fails, someone misses class. if the schedule is wrong, the app is useless. that is not a small UI bug.

i spent more time on clear errors. retries. ways to recover without starting over. screens that say what is going on.

a fix that took me an hour during thesis work took a day later. more users. more ways to break things.

it stopped being a school project for me. it became something people relied on.

the only question that mattered was: does it still work when things are messy.