4 ms·
Just scan the books, OCR the scans, and remove the copyrighted information that the publisher added to the codes. You then have a perfectly legal, digital copy
by skittles 14y ago
Just scan the books, OCR the scans, and remove the copyrighted information that the publisher added to the codes. You then have a perfectly legal, digital copy. Nobody does this because the codes change over time and there's no way to protect your work. It would be freely available to everyone as soon as the first copy hits an open server. It would also be nearly worthless as soon as the new version comes out (do you want to be the attorney that shows up to court without knowing the new changes?).
- officemonkey 14y agoSince passed laws are basically changesets of the existing code, it should be a simple matter of applying patches based on the new laws. In other words, once you digitize the code the first time, any git could update it. http://macwright.org/2013/02/13/the-code-compiled.html http://macwright.org/2013/02/13/the-code-compiled.html
- kyllo 14y agoNo pun intended, right? That might actually be a great idea--store all state and federal laws in public Git repos. Each bill introduced to the legislature is a pull request and if they pass it, it gets merged.
- officemonkey 14y agoIt would be amazing to read the entire history of the bill, from initial drafting through the congressional conference committee and see who was responsible for every change.
- kyllo 14y agoExactly. You could 'git blame' the law. Tech is so far ahead of government and industry it's mind-blowing. What is the D.C. Council paying Westlaw for, again? Having version control and a database?
- officemonkey 14y agoDon't forget burning a CD and printing off hard copies.
- bradleyjg 14y agoThe language to describe changes isn't fully standardized (at least for the US Code, I'm not familar with DC's legislative practices.) If you browse around it looks standard, but there are plenty of exceptions. So the best you can do is use a program to do a first pass.