[Kst] Image data type

[Barth Netterfield]

z(x,y) data presented as colour maps or as contour plots are really 2d 
objects, and are consistent with the current kst interface. We are 
implementing them now (hopefully working well by 1.0).

The 3d stuff (surface plots, C(x,y,z) surfaces, etc) are not at all consistent 
with the current kst2dplot interface, and will have to wait 'till post 1.0.  
As ponted out, they will almost certainly require significant upgrades, 
including using openGL, perhaps under other existing libraries.  Let's 
revisit 3D in the fall.

cbn

On July 22, 2004 04:53 am, Nicolas.Brisset at eurocopter.com wrote:
> On Wednesday 21 July 2004 23:37, Barth Netterfield wrote:
> > There are several ways to imagine '3d' data...
> >
> > A) Data on a regular grid, expressed as an array.
> >         (imagine 'kst -z 1 -nx 200 -ny 200 array.dat')
> > B) Data as a list of (x,y,z) points (as 3 arrays)  
> >         (imagine 'kst -x 1 -y 2 -z 3 map.dat')
>
> I would suggest extending that a bit to (x, y, z, color). I have developed
> in the past a small 'visual program' based on OpenDX (formerly IBM's
> DataExplorer, now Open Source) which took values resulting from a
> 2-dimensional parameter sweep (giving the x,y coordinates of grid points,
> not necessarily regular) and plotted a 3D surface using a third variable as
> z, with the color of the surface mapped from the value of a fourth
> variable. It turned out to be quite useful...
> We even ended up adding the ability to map on the surface two series of
> contour lines from a 5th and 6th variable as it allowed to visualize
> correlations between variables in a very compact way ! I guess that
> contrasts with what has been said yesterday (that contour plots are in
> essence 2D objects) as in that case they were actually mapped to the 3D
> surface. But maybe that can be achieved from a 2D object ?
>
> Well, I fear that sounds quite complicated. If anybody is interested, I
> could send the OpenDX 'program' I used (a sort of script actually) or even
> easier: pictures of the result. The thing has only 1 problem I never took
> enough time to solve: according to how (x,y) points are positioned
> sometimes the triangulation used in creating the surface connects points on
> opposite sides, which results in a very messy surface...
>
> > C) Data as a list of points, with an implicit pixel index (eg, healpix)
> >         (imagine 'kst -z 1 healpixmap.fits')
> > D) various representations of surfaces which are not functions (eg, the
> > surface of a sphere)
> >         (don't imagine anything - we're not going here!)
> >
> > -----------
> > Class (2) are openGL type objects, and will need a hightly modified mouse
> > interface, etc.
>
> I think there has already been some discussion to this in the past, but is
> it possible to reuse existing libraries from one of the following projects:
> - k/qmatplot (which is Qt-based and uses OpenGL for 3D, with quite nice
> classes I think)
> - scigraphica (gtk-based and maybe less easy to reuse, though I believe the
> product was pretty good until it was left unmaintained for a while)
> - LabPLot (also Qt-based) ?
>
> Then, there are also the very feature-rich libraries like VTK (which has
> nice Qt bindings I once saw somewhere on the web) or OpenDX which could
> certainly be reused...
>
> Just my thoughts,
>
> Nicolas
> _______________________________________________
> Kst mailing list
> Kst at kde.org
> https://mail.kde.org/mailman/listinfo/kst