I came up with a reasonable roadmap. Tell me what you think. These are
intended to be in chronological order:
goal 1 -
Read an image in, copy an RIO and write it out again.
This will need to utilize tile based processing.
We need to use only a small amount of memory (a little as enough memory
for a single tile, plus the app to as much as the whole image),
also configurable).
This will be used to develop cache, swap, and io stuff.
goal 2 -
Read an image in, rotate, and write out an ROI.
This will further develop the stuff above, as well as testing more
complicated graphs.
goal 3 -
Read an image in, convert to a differnt color space, write out that
converted colorspace, convert back, and write out the result. This
will develop color models.
------
Only when the above three operations are succsessful
goal 4 -
Write a current gimp plugin, utilizing this library.
goal 5 -
Fill out the operations
