I'm trying to talk Hover into implementing a simple DNS API. We have what we need from Amazon, but it's not a good idea to build on a single vendor's system. Having an alternative to swtich to if there's trouble would be good.
So here's how my outlining software uses the DNS API from Amazon. When outliners are used to author web content they often have a huge linkrot problem, because the address of an object is a function of its location in the outline. Move it, and all previous links broke.
But then we hit on the idea of placing a name on a headline, that stays with it as it moves around. And if that name were managed by DNS, we wouldn't even have to write a resolver for the name, DNS would do most of the work for us! Really neat when an ancient bit of technology ends up solving a very modern problem. Love it.
The important thing is that you didn't have to switch contexts, or load up your brain with a whole other lengthy task and then somehow try to regain the context of your work. You're much more likely, in teh first scenario, to avoid it -- too much trouble. In the second, once you understood what it did, it's relatively painless.