always@(*) statement

Don't use it for a clocked process, but for a combinatorial process, as you don't have to list the inputs that the outputs are based on. This task always caused problems whereby if you forget to add something to the sensitivity list, a latch can get inferred when you don't want it.

On a similar note, I'm a fan of the system verilog commands always_comb, which doesn't require a sensitivity list, and always_ff. Minimally, I think it keeps things clear. (I think there's supposed to be an error or warning if code the wrong thing, like a latch in an always_comb block, but I haven't really tested it out.