Whiteboard: ImagesDatatype	last revised by 127.0.0.1 on Aug 17, 2005 3:19 am

I believe it is going to be hard to apply xanalogical principles to images; the slicing and rearrangement operations common in text don't map well to two-dimensional, variable color-palette, infinite data formats.

Nonetheless, the following manipulations are needed for images:

1. extraction/reference of thumbnail mini-images

   the question is, if they are prepared in advance, at which sizes? and if not prepared in advance, computing them in advance would burden the server.

2. extraction/reference of a section of an image

3. combining of multiple images into one, e.g. for geographic maps.

Perhaps if we standardized on a lossless format (PNG) and a lossy one (JPEG) we could make it work, but I'm sure that won't please many people.

Hmmm, is there any research into segmenting/encoding of images, into fractal quadrants that might give us the ability to easily grab portions or scaled up/down pieces?