[Kst] [Bug 108084] New tool: differentiate curves

Barth Netterfield netterfield at physics.utoronto.ca
Thu Oct 20 19:38:57 CEST 2005


> ------- Additional Comments From arwalker sumusltd com  2005-10-20 19:33
> ------- I suspect that the problems with Curves indicates a problem in the
> underlying architecture. Would it not be better if we had the concept of a
> Curve and a Curve Instance. The former holds the vectors of the Curve,
> while the latter holds the properties (color, line style, etc) and knows
> how to draw a Curve. This way the user can have the same Curve with
> different properties. The Curve could hold default properties which would
> be used when a Curve Instance is created, essentially acting as sticky
> properties. There could also be the conpect of a copy of a Curve Instance
> in which modifying one Curve Instance would also impact the copies.

A curve is:
	1) an association of vectors
	2) the drawing properties of the association
Creating a new curve of the same vectors in order to have different properties 
is neither memory nor performance intensive, and leaving these two concepts 
is easier for the user in the common case, so I think we should leave them as 
they are.

> In the interim, I would propose that we simply cycle through the Curves. We
> could "mark" Curves as treated (as Nicolas suggests) but this represents
> more code and it doesn't really matter if we set the Curve properties based
> on the first or the last instance that we meet (which would be essentially
> random anyway).

I think that is fine.  If the user doesn't like it, the user can make a new 
curve with different properties.

> Incidentally, I'd like to change the two list boxes around in the
> "Monochrome Print Configuration" as the Available and Selected are opposite
> to the way they work in the Plot Dialog Content tab working on Curves.

I agree.



More information about the Kst mailing list