11 Jul
2008
11 Jul
'08
6:15 p.m.
While your reasoning is logically flawless, Martin's is IMO a much more useful behaviour. And making constant a property on normal typed variables (changing their variability to constantness instead), would be even better still.