Class Schedule Conflict Checker

Why a timetable hides its own collisions

A registration system will usually stop you booking two things in the same slot, and then it stops looking. What it rarely checks is the timetable you assembled out of one department’s lectures, another’s labs, a language class that runs fortnightly and a society meeting nobody put in the system at all. Those live in different places, arrive at different times of year, and the collision between them shows up in week three when you are standing in the wrong building.

This tool takes the whole thing as text, reads each line into a day, a start, a finish and an optional pattern of weeks, and then compares every pair. Pairwise comparison is dull but complete: with a dozen sessions it makes sixty-six checks, which is exactly sixty-six more than reading down the page tends to manage.

Touching is not overlapping

The single most common false alarm in software of this kind treats a shared boundary as a collision. A seminar finishing at 11:00 and a lecture beginning at 11:00 occupy no minute in common, so they are not reported. They may still be impossible — ten minutes of corridor between two buildings is a real constraint — but that is a travel problem with a different answer, and burying it inside a list of clashes makes the genuine ones harder to see.

Every session is therefore held as a half-open span: it includes its start and excludes its finish. Two spans conflict only when the later start falls strictly before the earlier finish. The check is one comparison, and getting it wrong by a single character turns every back-to-back pair on a full timetable into a warning.

The fortnight problem

Plenty of institutions run sessions on alternating weeks, labelled odd and even, or A and B, or by explicit week numbers on the departmental handout. Two of those can sit on precisely the same hours of the same weekday and never once meet, and a checker that ignores the pattern reports a clash that does not exist. The reverse is worse: a class every week against a class on odd weeks does collide, but only half the time, and that is the kind of thing a person discovers by turning up.

So week patterns are intersected before the clock is consulted. Odd against even yields nothing and the pair is dropped. Anything against a session that runs weekly yields the narrower of the two, and the result is reported with the weeks named, so a fortnightly collision is never quietly presented as a permanent one.

Sessions that run past midnight

An observing session, a radio show, a night shift in a teaching hospital: a finish time earlier than a start time means the session ran through midnight into the following day. Laying every session out on a single 10,080-minute week handles this without special cases, because the span simply continues across the boundary and is split where it crosses. A late Sunday session wraps round to Monday morning and can collide with the first thing there, which is a real result that a day-by-day comparison cannot produce at all.