A “codebase” is the entire collection of program text (aka “source code”) that is used to create some particular piece of software. The size of that codebase is usually measured in “lines of code”, much as you might measure the size of a book by counting the lines of text in it. The size of the codebase is reasonably well correlated with the effort that went into it, much like one book that's twice as long as another probably took roughly twice as much work to write.
This visualization shows the relative sizes of a number of different pieces of software, measured by lines of code. Note that the horizontal scale is different for the top group than for all the others. For non-programmers, I suspect there are quite a few big surprises here – things that perhaps looked easy to you turn out to be large codebases, or vice versa. For me, the big surprise was car software – I had no idea so much effort has been put into this.
But for everybody, the real kicker is at the bottom. Just look at how big the healthcare.gov web site (ObamaCare) software is! My very first thought upon seeing that: I'm not sure commercial efforts that large would ever produce something that worked – and there's no way in hell a government effort would!
Blow it up and start over, and make the first priority be this: how can we make this thing simpler? It's just too big to succeed...
No comments:
Post a Comment