In this case, I suspect that evaluation at the boundaries themselves is missing. The function I was plotting decreases monotonically with a large slope at the lower boundary, so the y-value at the lower boundary should be the highest value of the plot. I assume that the show boundaries overwrite the plot boundaries. I can see that this is necessary if a function has a vertical asymptote in the plotted range, but maybe it would be helpful to still make sure that the values at the boundaries are plotted. This is often what one is interested in and currently this impossible to plot. If the values at the boundaries tend to infinity, the user could simply move the boundaries a bit. What do you think, would this be useful/easy to implement?
Stan kcrisman wrote: > > On Jun 26, 9:47 am, Stan Schymanski <schym...@gmail.com> wrote: > >> This makes sense, thanks. The function that made me discover this >> behaviour does indeed tend towards an infinite slope at the boundary of >> the range I prescribed. However, I am not sure if I understand >> correctly. Can it happen that points outside of the range specified by >> xmin and xmax are evaluated? It does not seem to make sense to me to >> choose random points outside of the specified range. I thought that all >> the points chosen would include the xmin and xmax plus any points chosen >> in between. If a function gets evaluated outside of the specified range, >> it could lead to all sorts of problems if this function is not defined >> outside of this range. >> > > Insofar as I understand the plotting code, this should not happen - > and in fact quite a bit of work went into making sure that things plot > okay even if the function is not defined right at or near (say) your > lower bound. However, the plot boundaries are different from the show > xmin and xmax options, so maybe that is the confusion? But I don't > think what you are describing should happen; all randomization is > within the bounds prescribed e.g. plot(1/x,0,1) would randomize > between 0 and 1 but not plot outside that range, for the very good > reason you mention. > > - kcrisman > > > --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
