[Kst] Image data type

George Staikos staikos at kde.org
Thu Jul 22 22:05:17 CEST 2004

On Thursday 22 July 2004 10:51, Barth Netterfield wrote:
> Great.  Before too much gets written, does anyone have any comments on the
> class arrangement?  George?
> Thinking more about (B) (which is not happening today)
> Tri-linear interpolated makes sense for under-resolved maps, but may or may
> not make sense for over-resolved maps.  Then we *may* want to bin over a
> pixel rather than tri-linear interpolate to pixel centres.
>  This implies that we may have more kstArray sub-classes, like a
> kstBinnedArray.  Basically going down the kstEquationCurve, kstPSDCurve,
> kstHScurve path.

  I don't see anything wrong with that approach at this time, but it would 
really help to see a rough implementation or proof-of-concept.  As long as 
everything is relatively self-contained, inherits from the base classes, and 
follows the same design philosophy as the other classes (which it seems to) 
for consistency purposes, I think it will be fine.

George Staikos
KDE Developer				http://www.kde.org/
Staikos Computing Services Inc.		http://www.staikos.net/

More information about the Kst mailing list