I can think of quite a few todo list websites and apps off the top of my head, and the time-filling aspect looks like it’s heading towards the knapsack problem, so this idea is a bit too big and scary for me right now, I think. Your mention of priory assignment is cool, though, and reminded me of a chapter from Cryptonomicon by Neal Stephenson, which I don’t know whether you’ve read but is great.
I do know of lots of todo apps. I don’t know of any (without having looked) that let other people request additions to your list, or work with prioritization. Anyone can write a todo list app that sits on the phone and everything’s inputted by one person. It’s called a text editor.
I’m thinking of the people whose email inbox looks a lot like a todo list, because that’s how people ask them to do stuff. But that’s a kluge. I know these people exist because they’re usually complaining about it, and email clients are always trying to build in features to accommodate emails that are really todo-list items. What we really need is a new system, preferably one that integrates with the old one–maybe we should come up with a way to parse tasks from emails? That sounds complicated, but maybe it’s not impossible.
I get what you mean about the knapsack thing, but NP-complete problems aren’t generally difficult for small n. I’m talking about filling a few hours with a couple different tasks which might take 15min-1h. It gets way easier if your “specify how long this will take” function works in 15m increments. Then you’re working with, “I have 6 slots before my dentist appointment; find me 6 slots worth of stuff to do.” Also, if you don’t HAVE to fill every bit of time, the problem gets way easier. If the algorithm generally works, people won’t mind if it leaves 15m of free time or whatever–especially since the time estimates are necessarily estimates, and tasks often take longer than people think.
Also, if people have long stretches of time open, you generally want to make sure you get the longer tasks done then. If you have six hours of free time and doing taxes or something takes five hours, you need to prioritize that because big chunks of available time can be hard to come by. If you have a big swath of time and no long tasks, it’s a matter of sorting by priority. You decide which tasks to do to fill the spot, and then you prioritize them.
When it gets tricky is if you offer the user the option to split up big tasks. Say cleaning your kid’s room (which is a disaster) will take three hours. If you have a phone call scheduled early in the day and a hair appointment two hours later and you’re making lunch right after that and then your show is on an hour after that and the kitchen needs to be mopped sometime today… you’re probably not going to clean the room all in one sitting. Plus, you’d be pretty tired if you did it all at once anyway! So if you offer the option to break that up, then you’re approaching knapsack complications.
The solution to that would be to decide the smallest useful chunk of time to work in. 15 minutes is not really enough to pick up all the toys, and people tend to work better in slots of maybe an hour anyway. So if you split it up into 30m or 1h segments, then you’re back to relatively small n and you can work with that.
You also don’t have to use brute force–there are of course better algorithms for this–and you don’t have to get THE optimal solution. You just have to get one that works, and allow the intuitive human brain of the user to make changes to the suggested schedule.
This is a fun problem! Thoughts?