Mohtalim

research indicates that the number of bugs scales with the number of lines of code.

none sense, I ploted the number of bugs in my code vs number of lines of code and it's not linear. unless they are using log scale ...

I'm not sure what you are saying. Are you saying that for all the code you've ever written, you kept track of the number of bugs and the number of lines? Or are you saying that the number of bugs != the number of lines? Because I'll grant you that second one. But if you design something with more lines of code, it's going to have proportionally more bugs. That's what I was saying.

Here's another obvious, yet breaking news tidbit. Bugs exhibit locality (the technical term is clustering). Where one finds one bug, there are likely to be more nearby. Yes, this was a primary point in an AT&T paper, one of the ~80 papers I had to read in dependable embedded systems.

However, after hearing/experiencing some software engineering horror stories, I believe even the obvious stuff needs to be hammered in.

Dwindlehop wrote:I'm not sure what you are saying. Are you saying that for all the code you've ever written, you kept track of the number of bugs and the number of lines? Or are you saying that the number of bugs != the number of lines? Because I'll grant you that second one. But if you design something with more lines of code, it's going to have proportionally more bugs. That's what I was saying.

I was trying to make a joke about my bug/loc ratio is exponential

my bugs/line of code is linear

Yeah, I kinda thought it was something like that, what with the smilie. I was locked in to thinking (# bugs) < (# lines), though, and that's what threw me.

Buggy code is evil, but not as evil as bad requirements...