Ask HN: How do you make sure you're not reinventing the wheel?

I've often had some issue I thought I could fix with programming, like getting some task to automate before I knew about cron, or getting my OS X audio to not blow out my headphones if I left it high. I just worked for a few weeks in my lab on Matlab EEG processing scripts before discovering EEGLAB, a toolbox that does everything I've wanted to do. I had googled for existing solutions a little before each one, but couldn't find any of these. I like programming just fine but I'd rather be doing whatever it is I'm making this tool to do. How do you check to make sure something you want really doesn't exist?