Initial checks for unsolvable schedule

[russss]

The solver doesn't provide us with much help when it finds the schedule is unsolvable. We should run some initial checks on the scheduling problem before we hand it off to the solver.

Some things we could possibly check:

• Whether multiple talks from a single presenter are constrained to overlap - this bit us in 2024
• The "density" of talks by slot - this might help us find time periods where there are too many talks to schedule

[jellybob]

Also maybe worth checking if multiple events are scheduled to occur in the same venue at the same time, which I don’t think has bitten us before but feels like a good candidate for doing so.

[russss]

That's the "density" thing. There's nothing scheduled before the scheduler runs - but you could divide the number of slots used by a talk by the total number of slots available for it to be scheduled in, and then sum this up per-slot. If you've got a value of more than 1 for any slot then you probably have an issue.