Scope of Image in 2.2? (was: API Review RT:17407 Canvas Node)

[Martin Desruisseaux]

Hello all

Le 19/04/12 15:14, Kevin Rushforth a écrit :
> One thing we should have mentioned up front is that the ImgaeData (and 
> ImgFormat) classes are very preliminary, which is why they have started out 
> life as "@deprecated". The final form they will take depends on 
> http://javafx-jira.kenai.com/browse/RT-17409 .

Is it in the scope of the new Image API to support non-byte data? I'm thinking 
about oceanographer, meteorologist, astronomer, medical imagery and much more, 
where pixels may be floating point values, for example "/Sea Surface 
Temperature/" as real-world values in °C units, rendered to screen using a color 
palette.

To be more specific, something equivalent to java.awt.image.IndexColorModel 
would be critical for the rendering of "/Sea Surface Temperature/" maps for 
instance. Something equivalent to the ability to specify our own 
java.awt.image.ColorModel in Java2D would be nice for the rendering of floating 
point data, but less important than IndexColorModel (because custom ColorModel 
are slow in Java2D anyway, so we often transform the data into something more 
suitable to IndexColorModel. Nevertheless it was nice to have a ColorModel that 
can work directly on floating point data, even if slow).

It is okay if scientific imagery is considered out-of-scope for JavaFX (i.e. if 
JavaFX targets only vizualisation), since we can continue to use Java2D for that 
purpose - which work reasonably well - and convert to JavaFX only at rendering 
time. But some indications about the expected scope would be very useful...

     Regards,

         Martin