why one import path wasn't enough

after real students started using MySched, the next wall was not OCR quality. it was that one import path could not cover every school.

scanning a reg card got me through thesis. it did not get me through multi-school.

one path is easy to demo. then reality hits. every school looks different. some students only have a screenshot. some have a photo of paper. some can pick their section from a real list. some paste text from a PDF.

i could not keep forcing everyone through OCR.

so i built more than one way in.

if the school has a published schedule, pick school, term, section. clean data. no guessing.

if they only have a photo or paste, parse it on the phone. no account required for that.

if that fails, try AI. only then. and only when i allow it. that path costs more and needs more trust.

i keep the cheap correct path first. the fancy path last.

same reason i made accounts optional. if you need to sign up before you see your schedule, a lot of people leave. let them import first. sync later if they want.

once import was solid, the rest of the app made more sense. today view. reminders. tasks. widgets. none of that matters if the schedule is wrong.

the first feature got people in. better import is what made MySched feel like a real product to me.