Re: Throwing down the glove [was Re: Nanotech/Opensource: OpinionsNeeded.]

From: James Rogers (
Date: Mon Jul 17 2000 - 11:49:12 MDT

On Mon, 17 Jul 2000, you wrote:
> 80,000 lines in 3 months is about 900 lines a day. You'd have to hit the
> caffeine pretty hard, but it could be done. The 10 lines/day metric is for
> bug free code, and wraps up the entire software development lifecycle I
> think, at least coding+testing+maintenance, which justifies that tiny
> figure. Really, us coders are not such slobs that we can only manage 10
> lines of code a day!

Yes, the 10 lines/day metric includes all direct and indirect development
and support costs, so it doesn't really reflect coding speed per se. Note
that this type of accounting is how you apparently have the government
spending $400 on a hammer and similar; they include the direct and
indirect costs at all steps of acquisition/production rather than using
the "on paper" purchase price. While arguably more accurate, it isn't an
intuitive accounting methodology for the layman.

However, 80k lines of good code in 3 months does sound a bit far-fetched.
30k LoC in 8 months is very reasonable though.

> I bet that 30,000 or 80,000 lines of code was riddled with bugs!

It depends on the programmer and the environment. For example, I write
nearly bug-free code using C on UNIX at an extremely high number of
LoC/day, but tend to have much buggier code and program at a slower
rate if I do something in, say, Perl. For me, these two environments
pretty much represent both the best and worst case scenarios for
environments that I commonly use. C++, Java, and various 4GLs fall
somewhere in the middle. I'm sure there are programmers whose coding
preferences with respect to bugginess are exactly opposite of mine.

I have a suspicion that people who are really good at programming have
brains that are wired slightly differently than the rest of the
population. Every really good natural programmer I have known was able
to write nearly bug-free code very fast in their preferred environment.
This doesn't actually speak for the efficiency of the environment, only
for the particular way a programmer is wired. However, in my experience
the majority of people working as programmers don't have an innate
cognitive ability with respect to programming and tend to write relatively
buggy code in any environment by comparison. Every successful development
team seems to have at least one "natural" programmer on board, and a
disproportionate percentage of the final codebase is typically written by
these individuals.

-James Rogers

This archive was generated by hypermail 2b29 : Mon Oct 02 2000 - 17:34:49 MDT