Why MySched needed more than a scanner

The first version of MySched scanned a registration card and turned it into a schedule. That worked for my thesis, but students at other schools didn't always have the same kind of input.

Some had screenshots. Some had photos of paper. Others could choose their section from a published list or paste text from a PDF. Better text recognition could improve the scanner, but it couldn't make one input method fit all of them.

So I added more ways to import. If a schedule is already available in the app, a student can select their school, term, and section. If they have an image or text, they can start there instead.

The order mattered too. If a published schedule already gives the app correct data, there's no reason to pay for AI to interpret it. AI can help when more interpretation is needed, but I have to consider the cost and how the input is handled.

I also kept accounts optional. I wanted students to find out whether the app could handle their schedule before asking them to sign up. Sync could come later.

This mattered beyond the import screen. Today view, reminders, tasks, and widgets all start with the schedule. I could keep adding features, but students still needed a way to get their actual timetable into the app correctly.