On Mon, Sep 28, 2009 at 3:28 PM, Andrew Coppin <andrewcop...@btinternet.com> wrote: > Ben Franksen wrote: >> >> If it turns out that a >> precise spec is unwieldy (too complex) then this is a hint that maybe it >> is >> not a good abstraction. > > Or your specification language is insufficient to describe it... > > (I don't know about anybody else, but I find that when I use QC, about 75% > of the bugs reported are bugs in the spec, and only 25% are bugs in the > thing I'm actually trying to test...)
Maybe that means that you don't apply the same level of care to testing as you do to coding your main program. Or perhaps that is the quality of many bugs: the code does exactly what I intended, but what I intended was wrong. Luke _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
