Open twneale opened 12 years ago
Wow, I'm 3 months late responding to this (sorry @konklone), but yes, I'm in favor of as much documentation, comments, and unit tests as is practical. People will be really critical of this data--that much we count on--so ideally we'd want to follow a defensible process. I'll keep these things in mind as I keep working.
I think Andrew was right, that we really just need to publish the whole thing at once. Let's just get this into a format we can start to use and understand (maybe splitting it into manageable sized chunks and run through a pretty-print formatter?) and start analyzing its accuracy and comprehensiveness. I think we are dangerously close to publishing a JSON version of the US Code as a flagship dataset.
Do you think some extra documentation/comments, or unit tests, would be appropriate in shoring up the code?
If it's easy to do, I do still think there'll be some value in extracting a structure-only form of the US Code -- but since Andrew's correct in that we're going to want the content anyway and we'll have to parse that, it'd mostly be for public use, so maybe I'll try doing that if I feel like it - otherwise let's not spend effort on it that we don't need for our own work right now.