Here’s my contribution to the lore of software development:
Any piece of software larger than a screenful is a steaming pile of crap.
Pessimistic, yes, but perhaps also liberating in a way. Think about it.
Update:
Apparently I’m not the only one with this sentiment.
Knox’s conjecture:
“Any piece of software smaller than one screen won’t do much interesting. Also the code might suck anyhow.”
Oh well.
Haha, yes John I think your conjecture is a good one.
Really, though, I would love to hear about large high-quality code bases. I am starting to think they don’t exist.
that really depends on your definition of “large” and “high quality”.
MediaWiki the software behind Wikipedia is many many lines of code, and i’d say the resulting product is quite high quality.
I think it comes down to human nature when you think about it. It’s human nature to want to lead; as such, the larger the project, the more people think it should do X instead of Y. Thats what leads to shit code.
Your law is just a different interpretation of the old “too many cooks spoil the broth” adage.
I’ve browsed through MediaWiki’s code. It’s a pile of crap. I mean, it works, but it’s very badly written (PHP, what can you expect).
However, large high-quality code softwares do exist. I can vaguely recall being very impressed with how well written Apache is (I mean, for a C program…). I also think “distorm” is well-written (again, for a C program. Also note he is my friend, so I may be biased).
You have to remember that most open-source is partly (and sometimes completely) written by amatuers. How much professional code did you get to read?