In a plot draw a function (for example y=x^2), then choose dashed line as line style. A solid line will be draw, except you choose very few "points" in the interpolation section. But it should also work with many points. A relevant question is also: would it be posible to choose some interpolation styles between the points (may be also usefull for svg rendering)?
(In reply to H.H. from comment #0) > In a plot draw a function (for example y=x^2), then choose dashed line as > line style. > > A solid line will be draw, except you choose very few "points" in the > interpolation section. But it should also work with many points. The line is drawn between the data points. If you have many data points that are plotted close to each other, the actual style of the line (solid, dotted, etc.) cannot be resolved. If you zoom in you should see the difference. For many data points the line style or drawing the line at all is not relevant because this cannot be resolved anyway. To be able to resolve we'd need to drop many data points, i.e. don't plot them. This can work for some well-behaved functions like x^2 where not so many data points are required to get a reasonable results. But for noisy data withe many outliers it's not immediately clear which data points to drop. If the user requests to plot 10k data points on a small area, wants to really see all of them and wants to clearly see the dotted line, here I don't know how this should work... > A relevant question is also: would it be posible to choose some > interpolation styles between the points (may be also usefull for svg > rendering)? We have different interpolation methods. Click on the xy-curve, go to in the properties explorer to the tab "Line" and select one of the interpolation methods under "Type". Is this what you're looking for?
This is not optimal, because if you plot math-functions (not from spreadsheets), it makes no sense to have different data-points at all (it is a labplot-technical implementation-detail, which in other case could be hidden from the user - in this case it is useful for me to reduce the points to show dashed lines). But I understand that it may be technically difficult to implement.
(In reply to H.H. from comment #2) > This is not optimal, because if you plot math-functions (not from > spreadsheets), it makes no sense to have different data-points at all (it is > a labplot-technical implementation-detail, which in other case could be > hidden from the user - in this case it is useful for me to reduce the points > to show dashed lines). But I understand that it may be technically difficult > to implement. I agree with this. The number of points we ask the user now to provide in this case is a technical implementation details and it's only because we don't have a more advanced algorithm for this. We should improve here. Let's keep this bug report open. I'll check how kmplot is doing this.