Wishing we had COLORRAMPs or something

[Pierre Racine]

Stephen,

This is EXACTLY what I would need right now. This would help making
MapServer a much better cartographic product, event better than ArcMap. It
is harrassing to define so many CLASSes right now, mostly when we want to
assign SIZE and COLOR to a symbol. A COLORRAMP would do the job perfectly.

Is this going to be available soon?

We also need to fix bug #1398 to allow SIZEITEM to control the size of
point data type. It would be great to allow the construction of SIZEBREAK
similar to what you described for COLORRAMP.
To define COLORRAMP we would have:

    COLORRANGE 0 0 0  255 255 0  255 0 0 #black to yellow to red
    UNDERCOLOR 0 255 0 #green
    OVERCOLOR 0 0 255 #blue
    DATARANGE 0.0 100.0 #with this range, -10..0 would be mapped to green,
0.0001 to black, 50.0 to yellow, 100.0 to red and 100.0001... to blue
    RANGEITEM "foobar"

Or

    COLORBREAK 0 0 0  255 255 0  255 0 0 #black to yellow to red
    UNDERCOLOR 0 255 0 #green
    OVERCOLOR 0 0 255 #blue
    DATABREAK 0.0 30.0 90.0 100.0 #instead of DATARANGE. The number of
break-1 must match the number of COLORBREAK. x<=0.0 = green, x<=30.0 =
black, x<=90.0 = yellow, x<=100.0 = red and x>100.0 = blue
    RANGEITEM "foobar"

In this last example we could use COLORRANGE instead of COLORBREAK. This
would imply that the values between 0.0 and 30.0 would be mapped to a color
between black and yellow. This makes clear the difference betweem using a
RANGE and a BREAK. A range implies interpolation and a BREAK implies exact
color. In this case the number of color must match the number of breaks.
Example:

    COLORRANGE 0 0 0  255 255 0  255 0 0  0 255 255 #black to yellow to red
to cyan
    UNDERCOLOR 0 255 0 #green
    OVERCOLOR 0 0 255 #blue
    DATABREAK 0.0 30.0 90.0 100.0 #The number of break must match the
number of color in COLORRANGE. x<=0.0 = green, x<=30.0 = interpolation
between black and yellow, x<=90.0 = interpolation between yellow and blue,
x<=100.0 = interpolation between red and cyan, x>100.0 = blue
    RANGEITEM "foobar"

________________________________________________
Similarly for SIZE we would have something like:

    SIZERANGE 1 100
    UNDERSIZE 1 #omitting this parameter would default undersized value to
the first value of SIZERANGE
    OVERSIZE  100 #idem to the last value
    DATARANGE 0.0 1456.0 #this would keep the determination of symbology in
the map file instead of having to compute a value for SIZEITEM in Postgres
for example. (which is the data side. "Keep data and their representation
separated")
    SIZEITEM "foobar"

Or with breaks:

    SIZEBREAK 1 10 100
    UNDERSIZE 0
    OVERSIZE  100
    DATABREAK 0.0 300.0 900.0 1456.0 #instead of DATARANGE. The number of
break-1 must match the number of SIZEBREAKs. x<=0.0 = 0, x<=30.0 = 1,
x<=90.0 = 10, x<=100.0 = 100 and x>100.0 = 100
    SIZEITEM "foobar"

Or

    SIZERANGE 1 10 20 100
    UNDERSIZE 0
    OVERSIZE  100
    DATABREAK 0.0 300.0 900.0 1456.0 #instead of DATARANGE. The number of
break must match the number of SIZEBREAKs. Value 150.0 is mapped to 4.5,
value 900 to 20, etc...
    SIZEITEM "foobar"

It is important to make the distinction between RANGEs and BREAKs.

Pierre Racine