Development principles

I'm just collecting these.  Other, far smarter people put 'em together -- but I'm going to compile them somewhere so I can later say to myself, "Self, you dummy, that's a principle that you not only know, you wrote it down."  Then link back to it.  Feel free to contribute.

• Say no to "do-overs."
• You have a big codebase; knowing what you know now, of course, you want to rewrite it all... but while your memory tells you 90% of it would benefit from the rewrite, something more like 20 - 40% is probably an obscure, difficult-to-remember bug fix or work-around or similar, and you'd end up making something pretty -- that broke all over again.
• Attributed to ....  (I'll have to look that up.)
•