[Kst] Image data type

[Barth Netterfield]

On July 21, 2004 03:35 pm, Rick Chern wrote:
> Barth Netterfield wrote:
> >Schedule:
> >I am *very* hesitant to start down this path now.  There is a lot to
> > consider and get right, and I would much rather have us focusing on a
> > high quality 1.0 release!  But.... if you are sure you can do this
> > without jeprodizing 1.0.... perhaps (gulp!)  Can you have something
> > mainly working in 1 week?
>
> Actually, I am not sure how long it will take, but would it be okay if I
> hid the parts of the user interface relating to images before the
> release, if it is not complete (mainly just removing the "Image" button
> from the data manager, and altering the plot dialog)?  Currently I am
> able to create images through the data manager, and add them to plots
> using the plot dialog.  However, I am just starting on the plotting of
> images, and I agree there is a lot to think about and work out.

I guess I need an accurate estimate as to how long it is going to take to make 
solid.  Lets finalize the API in any case

> >On July 20, 2004 02:00 pm, Rick Chern wrote:
> >>I am currently working on an Image data type.  Description of images:
> >>- images will be specified by three vectors, representing x, y, and z
> >
> >So you intend to support un-gridded data.... What is your binning policy
> > going to be?  What do you do about over-sampled maps?  about undersampled
> > maps? How will you decide on gid density?
>
> I am not sure yet - in fact I think I may work with gridded data first,
> and see how that goes.
>
> >>- it is up to the user to make sure the three vectors selected make
> >>sense as an image
> >
> >What does this mean?
>
> For now, I think this just means the three vectors need to be the same
> length.

This needs some more thought: 
 
A regular grid is a 2d array: Z(X0+i*dx, Y0+i*dy) - this is not efficiently 
represeted by 3 vectors: Maybe:
KstImage(KstVector Z, int nx, int ny, double dx, double dy, double X0, double 
Y0) 
- ie, 1 vector and a bunch of numbers.  3 Vectors implies that it is not on a 
grid.

BUT: the KstImage should present to kst2dplot the same interface, regardless 
of the nature of the input data.

I thought sent out a note on this a while ago, but can't find it.  Does anyone 
else have it?  Am I imagining things?


> >>- images will be plotted on the 2D plot, using colours/shades to
> >> represent z - only 1 image can be plotted per plot
> >
> >What will the colour-scale interface be?  Also, one might want to show
> >multiple images covering different parts of the XY space, or to have a
> > colour map under a contour plot.  So I wouldn't limit to one image per
> > plot.
>
> Would contour plots be a separate data object from colour maps?  I think
> I will focus on getting 1 colour map plotted correctly first, and then
> add the ability to have multiple maps.

No.  In the same way as curves have 'hasLines' and 'hasPoints', etc, Images 
would have 'hasColormap' and 'hasContourmap' both of which could be true.

>
> _______________________________________________
> Kst mailing list
> Kst at kde.org
> https://mail.kde.org/mailman/listinfo/kst