I am interested in creating a forum/blog where we can discuss various ways to structure tasks for part of a problem, together with suitable programming. Perhaps a "better than this" or "instead of this" blog would be useful --- what do you think? [FWiW: I tilt toward the latter because slick stuff is not always optimal for learning.]

In the meantime, consider various ways to generate a random polynomial (without structural constraints). Here is my current code for a problem where y-intercept is the only feature of this function needed by a student [full problem involves other stuff]. I request suggestions for improvements or slicker ways to generate a random polynomial.

#### integer coefficients (upto sign)

@a = (1 .. 9)[NchooseK(9,4)] ;

#### integer powers for non-constant terms

@p = (1 .. 9)[NchooseK(9,3)] ;

#### powers WITH a constant term

@p = ( 0 , @p )[NchooseK(4,4)] ;

$poly = '0' ;

foreach my $i (0 .. 3) {

my $s = list_random(-1,1);

$poly .= "+ ($s) * $a[$i] * x^($p[$i])" ;

} ;

$poly = Compute( "$poly" ) -> reduce ;

$y0 = $poly->eval( x => 0 ) ;

?? Q for Davide: should reduce be reduce()

## inhibit rearrangement

Context() -> reduction -> set( '(-x)+y' => 0, '(-x)-y' => 0 ) ;

## repeat reduction of $poly to remove parentheses

## around constant if it is the last term

$poly = Compute( "$poly" ) -> reduce ;

I suspect the first use of reduce would replace "+ (-5 x^0)" with "+ (-5)" and the second reduce yields "- 5". It is unclear why the second step is not needed if the constant is not the last term.