Who would’ve thought that if you encode, you must also at some point decode.
A little while ago I implemented the logic for turning native Urbit data types into something Ethereum nodes can read. I got distracted for a good little while by a vaguely related requirement, and am only now getting back to the thing I was originally working on. And now I need to add decoding to my encoding.
The logic for it all isn’t actually that complex, having to dive back into it (but this time, doing it the other way around) isn’t that big a deal. A couple minutes reading the spec again brought me back up to speed. But it’s an inefficiency I could’ve avoided, and I’m sure the whole implementation I’m working on now could’ve been prettier faster if I still had my head in the related logic.
Hindsight 20/20 though. My foresight, well, sometimes I feel like I may need laser correction.