[Haskell-cafe] RE: Answers to Exercises in Craft of FP

[Hamilton Richards]

Dear Chris--

Many of us instructors who use (or have used) these textbooks (or 
others that have exercises) in university classes have found from 
experience that

1. Students learn best from exercises when they make a real effort to 
solve them before looking at the instructors' solutions.

b. Students tend not to attempt exercises that don't count towards 
their grade. They don't usually decide consciously not to do the 
exercises, but if the exercises don't count, no deadline means 
anything, and the exercises get put off in favor of homework in other 
classes that actually counts. (I've found a "sweet spot" at 10%-- 
that's enough to induce most students to hand in most of the 
exercises, while minimizing the advantage of those who merely put 
their names on a copy of someone else's paper.)

iii. If homework counts and solutions are available before the 
hand-in deadline, some students will find those solutions 
(particularly easily if they're available on the web) and hand them 
in as their own. They thus gain an advantage (I like to think it's 
only a temporary one) over students who play by the rules. It seems 
to me that a policy that makes folks who follow the rules feel like 
suckers is a not a good policy.

An instructor who is teaching a course for the first time has plenty 
of work just organizing lectures, selecting exercises, overseeing 
their grading, and making up and grading exam questions. After a 
semester or two, when the lectures have stabilized, one typically 
starts making up new homework exercises. The first time around, 
however, it's really nice to have exercises that have already been 
devised, and used in actual classes, by the textbook's author.

So I don't believe your misery is the result of a deliberate plot to 
make your life hard. It's the result of instructors organizing things 
for the benefit of their students, and authors and publishers 
addressing the needs of their largest market.

In my own classes, I hand out suggested homework solutions, but only 
on paper (although they could be scanned and put on the web, I 
haven't yet seen that happen). Exam solutions, however, I make 
available on the web, because all the exam questions are composed by 
me, and used just once (or used again only after a lapse of quite a 
few years). I currently use Haskell in my Programming Languages 
class, and five semesters' worth of exam questions (not all Haskell, 
of course) with solutions are posted on-line at 
http://www.cs.utexas.edu/users/ham/UTCS/CS345/OldTests/cs345OldTests.html 
. These aren't the same as textbook exercise solutions (the context 
is not as explicit as it is for exercises at the end of a chapter), 
but you're welcome to them.

If you use these up --or don't find them useful-- you might reflect 
on the fact that the conference papers and journal articles from 
which many folks learn lots of "interesting stuff" seldom include 
exercises.

With best wishes,

--Ham

At 11:37 PM +0100 2005/1/26, Christian Hofer wrote:
>Maybe I am too much rooted in the German university system, where 
>the students' autonomy is held high (against all evidence). But I 
>never understood, why we - who have to learn the interesting stuff 
>completely on our own, because bad luck supplies us only with Java 
>teachers (although other professors use Scheme, Lisp, Prolog) - are 
>punished for a possible misuse of the exercises' solutions by 
>students who would by cheating loose some years of their lifetime 
>before failing in the exams.
>
>Regards,
>Chris
>
>Am 26.01.2005 um 15:41 schrieb Paul Hudak:
>
>>I'm not sure how Simon Thompson feels, or other instructors using 
>>his or my book, but a downside of posting all of the solutions is 
>>that the problems cannot be assigned for homework.  I have many of 
>>the solutions to SOE problems, and could post them, but am 
>>wondering if it would be better to make them available only to 
>>instructors, or to those who might convince me that they are not 
>>reading the book for credit.  Or is that being too draconian?
>>
>>   -Paul

-- 
------------------------------------------------------------------
Hamilton Richards, PhD           Department of Computer Sciences
Senior Lecturer                  The University of Texas at Austin
512-471-9525                     1 University Station C0500
Taylor Hall 5.138                Austin, Texas 78712-0233
ham at cs.utexas.edu                hrichrds at swbell.net
http://www.cs.utexas.edu/users/ham/richards
------------------------------------------------------------------