My Brain Hurts
This has been a very theory-intensive day. So far I have:
- Explained the proxy-stub design pattern to Leah.
- Explained to a user why
interface I { }
class K where T : I, P> { }
class A { }
class B : I { }
class M : K { }
does not compile. (Clearly B cannot be used as T in K in the last line. Why? B is an I, but T is constrained to implement I,A>. The fact that an M is an K is irrelevant -- generics are neither covariant nor contravariant in C#! There can be no is-a relationship between I and I when N extends M)
- Described to a curious coworker the departures of large real-world datasets of decimal integer, decimal float and hex integer most-significant-digit distributions from predicted logarithmic (ie, Benford's Law) distributions.
- Deduced a correct (I hope) design for the definite assignment algorithm for the null coalescing operator -- which, mysteriously was never specified in the C# 2.0 specification.
- Wrote three lines of production code implementing said algorithm:
ScanExpr(expr->exprTest, bsetCur, bsetErr);
BitSet bsetTemp(bsetCur, &GetHeap());
ScanExpr(expr->exprElse, bsetTemp, bsetErr);
Wow, do I feel productive! Three lines of code in only 7 hours. Where does the time go?