Blend function design somewhat impractical

The new features are absolutely amazing! That curve editor is really a nice piece of work :-S

I have tested the new scripting features and they seem to work well. However one thing about the blend function design is really annoying: why have a separate function for *each* blendmode and not just an index parameter?

Take a look at the code here:


.. with a simple index the above 11 lines conditional could have been skipped and moreover all blendmodes would be available:

Code
blend(r1,g1,b1,a1, r2,g2,b2,a2, opacity, hdr, BL_MODE)

You could preregister constants for the different modes (BL_NORMAL = 1 etc). If no index is supplied, 1 could be assumed.

(and please please.. don't forget the script editor in V3, it is very annoying that you can't detach/resize/open in new window.. I bet your own programmers did not use it when making example scripts..)

Please FF - if you do not revise the blending function design, could you at least create separate HDR function definitions from non-HDR definitions..

The problem is that the parameter count differs btwn. non-HDR functions and HDR functions, which makes it very impractical.

Consider the following alternative (and much cleaner) script of your blending script example:

Code

function prepare()    BL_MODE = get_intslider_input(BLENDING_MODE)    HDR =  (1 == get_checkbox_input(HDR))    BLENDFUNCTIONS = {       blend_normal,       blend_darken,       blend_multiply,       blend_color_burn,       blend_linear_burn,       blend_lighten,       blend_screen,       blend_color_dodge,       blend_linear_dodge,       blend_overlay,       blend_soft_light,       blend_hard_light,       blend_vivid_light,       blend_linear_light,       blend_difference,       blend_hue,       blend_saturation,       blend_color,       blend_luminosity    } end; function get_sample(x, y)    local r, g, b, a = get_sample_map(x, y, BACKGROUND)    local r2, g2, b2, a2 = get_sample_map(x, y, FOREGROUND)    local opacity = get_sample_grayscale(x, y, OPACITY)    return  BLENDFUNCTIONS[BL_MODE](r, g, b, a, r2, g2, b2, a2, opacity, HDR) end;

Now this a much better alternative compared to the cluttering and inefficient conditional check seen in:


But the darn parameter inconsistency caused by the additional HDR parameter present in some modes, causes the script to either fail on non-HDR modes (as presented above) or fail on HDR modes...

I suggest not mixing things here, so the HDR modes are available as separate functions (since these are really different), i.e.:

blend_normal_hdr
blend_darken_hdr
blend_multiply_hdr
blend_lighten_hdr
blend_linear_dodge_hdr
blend_difference_hdr
blend_hue_hdr
blend_saturation_hdr
blend_color_hdr
blend_luminosity_hdr