In the image below the top curve is a spline. You can see its 5 fit points (square grips). The middle curve is the spline converted to a polyline with splinedit p (resolution 10). The bottom curve is the middle curve converted to a 2D polyline with pedit s (spline fit option).

Zooming into the yellow box of the spline we get a smooth curve. In fact the curve only looks smooth. The display of the equation for the spline is being evaluated at very small intervals that are so small you do not see that curve is being drawn as a series of straight lines that are then converted to pixels.

Zooming in on the second curve yields the following which obviously show the polyline approximation of the smooth spline.

If pedit s(pline) is used with this polyline as @Kent1Cooper mentions a new 2D polyline will be created that is similar to the original source polyline. Zooming in on it shows that it creates a smooth curve between the vertices of the polyline. In this highly zoomed in view you can see the bright white curve of the splined polyline, the dim white lines of the polyline, and the original spline (red).

It is important to note that no matter how a curve is represented, whether a polyline with a series of straight and arc segments or a spline, the NC driven waterjet will drive the jet in short, straight line increments. You would probably not notice the difference if the polyline you have is used or if it were represented as a smooth spline.
What is your goal in making it a "simple curve"? Does the water jet's post processor handle splines?
Note, a good strategy in getting a smooth spline is to minimize the number of fit points. Having too many fit points increases the likelihood of getting ripples in the curve.
lee.minardi