Creating ideas is easy. Shooting them down is the real skill.
A long time ago I did some work on a game called Zombie Run. Players ran around in the real world, tracked using their GPS, while invisible enemies closed on them. I had plans to build it out further, creating better maps, smarter enemies, levelling and progression. Early playtests were a lot of fun. Everything looked good.
And then I realised that what I'd invented was the playground game Chasey - for people with no friends.
Why put all this technology, all this work into creating artificial people to chase you around? Why not just use real people?
If the problem is "make exercise fun by adding an escape mechanic", then Zombie Run does solve the problem. But Fire Hazard solves it better. Just because you can solve a problem with code doesn't mean you should. I scrapped Zombie Run and ran a real-world games weekend called Camp Fire.
More recently, I've been working on something called Relentless Memory. Memory decays exponentially over time, but if you prompt yourself at regular intervals, you get much better at learning. Relentless Memory displays segments of my notes and kindle highlights using spaced repetition.
The problem is that I have to make a point of going to relentlessmemory.com and doing the review, using prime productive time on my laptop. I'd rather use my commute, where my mind is mostly idle (smashing through Old St traffic doesn't take much concentration these days). Right now I listen to podcasts of new material. What if, instead, I listened to automatically-generated compilations of my notes from previous learning?
This kicked off a project called Memorycast. The plan was that when I added notes to Relentless Memory I'd also record them in my voice. The software would run the spaced repetition algorithm, collate the recordings, and deliver them as a podcast each morning for the commute.
I'd be regularly reviewing my notes, as well as reading them out and explaining them to myself (vocalising an idea forces you to check that you really understand and accept it). I wrote some prototype code, considered entering Evernote's Devcup competition, and started to sketch out the commercial potential.
And then I realised that what I'd invented was book clubs - for people with no friends.
Screw that. It'll work and it'll work well but it's the wrong way to solve the problem.
I abandoned work on the code and invited some friends around for dinner instead.