Should a consistant means of specifying an image region be used?

[robert.desonia@…]

I noticed that functions like psImageSubset specifies a region using a
lower-left corner (x0,y0) and size nx & ny. Functions like psImageTrim,
however, specifies a region using a lower-left corner (x0,y0) and a
upper-right corner (x1,y1). psImageSubSection basically uses this second
method, though the order of the elements are different ( [x0:x1,y0:y1] versus
x0,y0 - x1,y1 ).

A single method of describing an image region probably should be adopted for
the library.

-rdd

[eugene]

the concensus was (x0,y0) - (x1,y1) where (x1,y1) is exclusive allowing the use
of (0,0) - (Nx,Ny). the particular string parsing function is needed since this
is a standard FITS representation of a region.

APIs need to be inspected for consistency.

[Paul Price]

psImageSubset, psImageTrim, psImageReadSection, psImageSlice are the ones that
require image sections to be defined. Each has been brought into line with the
standard x0,y0,x1,y1 region specification. Note that x1,y1 are exclusive, and
may be negative (in which case they are relative to the particular dimension).

[Paul Price]

Should be fixed in SDRS-07 and ADD-06 (7 September 2004).