Successfully architected solutions do two things: First, they rely on existing open standards rather than reinventing the wheel. They rely on some of the internet’s greatest hits, things like OAuth and REST, and store data in formats born in the internet age, formats like GeoJSON and markdown. No licenses, no SDKs, just data. Second, they’re built as a dumb core with a smart edge. Upgrading a standard is a monumental task. Upgrading a tool is trivial. But more importantly, there’s room at the edge for experimentation, and with readily available libraries, amazing vehicles of empowerment like geojson.io, something that nobody knew could exist six months ago, suddenly start appearing over night.
First, go read the article, it’s really good and packed full of interesting points. I’ll wait.
Now imagine the CFR stored as data on GitHub.1» A GitCFR repository would be open to all and exposed to the APIs of GitHub. Besides using GeoJSON to locate fire hydrants in your neighborhood you could use GitCFR to find regulations relevant to the manufacture of those fire hydrants. Any API call would return just a specific piece of the regs that could be displayed as the app builder desires.
Of course any of this would require that the GPO set up a system for loading the CFR to GitHub so we don’t have to worry about issues of authenticity. While anyone can grab the bulk XML of the CFR from the GPO’s FDsys website and load it into GitHub, it really needs to the be done by the GPO so that developers can rely on the authenticity of the data.2»