+Build gamma tables using fixed point (and do away with floating point entirely).
+Avoid building gamma tables whenever possible.
+Use greater precision when changing to linear gamma for compositing against
+ background and doing rgb-to-gray transformation.
+Investigate pre-incremented loop counters and other loop constructions.
+Add interpolated method of handling interlacing.
+Provide for conditional compilation of 16-bit support (except for the
+ initial stripping down to 8-bits when reading a 16-bit PNG datastream).
+Switch to the simpler zlib (zlib/libpng) license if legally possible.
+
+*/