plot2d(1/x,[x,0,1],[y,0,5]); => division by 0

>>>>> "Felix" == Felix E Klee <felix.klee.maxima@gmx.net> writes:

    Felix> On Thursday 17 April 2003 17:02, Raymond Toy wrote:
    >> None that I know of.  All you can do right now is to change [x,0,1] to
    >> [x,eps,1] for some suitably small non-zero eps.
    >> 
    >> We could, however, change the plotting routine so that the end-points
    >> are never evaluated.  That's what quadpack does.

    Felix> Surely this would work if the singularities are at the endpoints. But what 
    Felix> if plot2d steps on singularities in between? To solve this problem, I 
    Felix> recommend the way Mathematica handles it: Plot just ignores values it can't 
    Felix> evaluate and returns warnings instead. Unfortunately I couldn't come up 
    Felix> with an example that deals with singularities, but I was able to forge an 

plot2d(1/(x-0.5),[x,0,1]) should demonstrate this, but the adaptive
plotting is taking forever to compute the points.  It's slowly moving
towards 0.5.

I think we really need a way to turn off adaptive plotting.

Ray