8 comments

[ 334 ms ] story [ 121 ms ] thread
When I hear statements like 2 million lines of code, I ask myself whether this is actual code that a team of developers wrote or if they're counting the lines in third party packages and frameworks they are using. LOC is not an accurate measurement of the quality of a software application. Interestingly it seems that it is written in the Ada programming language as well (a favourite amongst mission critical software applications and military purposes): http://en.wikipedia.org/wiki/Ada_%28programming_language%29
I'm tempted to read it as "merely" 2 million given how bloated these projects tend to become. healthcare.gov is reported to be 500 million LOC, in comparison. I simply don't know how one writes that many lines, but it happened.

(And FWIW, I'd feel much safer if the figure for ERAM were 200KLOC or 20KLOC.)

> I simply don't know how one writes that many lines ...

I think with a template get ide and a verbose language ... Think eclipse and Java ...

LOC is an accurate measure of quality, but it's an inverse measure.
Rubbish. LOC is a measure of size. How well it is put together is entirely something else. Seeing that it is done in ada, at least it shows the system architects have a clue about large software systems.
So the Linux kernel is worse quality now than in 1991?
100 developers @ 10 LOC/day @ 200 day/year @ 10 years = 2,000,000 LOC.