[plt-dev] random testing, we need to do more of this

From: Michael Sperber (sperber at deinprogramm.de)
Date: Mon Jan 25 09:55:59 EST 2010

Carl Eastlund <carl.eastlund at gmail.com> writes:

> On Mon, Jan 25, 2010 at 9:45 AM, Michael Sperber
> <sperber at deinprogramm.de> wrote:
>> Sorry if I mischaracterized - FastTest does have higher-order
>> generators, but no generator datatype, so compositionality is limited.
>> The most useful set of generators that's missing is what I mentioned -
>> generators for procedures.  (And these, as far as I know, can't be
>> constructed from simple generator procedures, as used by FastTest.)
> The design of FastTest is probably somewhat quirky, due to initially
> being motivated by DoubleCheck, the ACL2 version, which is necessarily
> first-order.  Nevertheless -- I'm not sure what you want from a higher
> order generator.  If I wanted a random function from strings to
> integers, I'd probably write (lambda (x) (random-integer)).  

That doesn't appear to implement a function to me :-)

> If you have more experience with these things, please let me know what
> else is needed.

I keep having that conversation with people who're curious about
QuickCheck. (I've so far had it with Peter Thiemann, Matthias, Torsten
Grust, and others.)  All I can say is that my explanation isn't likely
to be more effective than the original QuickCheck paper.

Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla

Posted on the dev mailing list.