TODO:
-Compute pullups
	FIXME: the computation is suppose to be in there
-Remove pullups transistors?
	Maybe add an option whether to omit them or not
-Give a .txt input file that specifies a point to node name
	(assumes each node has at least one unique spot which seems reasonable, there should be at least one via)
	Instead I'm doing it as layer labels
-Generate transistor polygons
	FIXME: using the whole poly as the transistor
-Add support for p-mos
	Needed for 4004 and friends
-Add command line options
-Smooth same layer polygons if points are very close
	Uncessarily increases rendering complexity
-Be able to parse all layers from a single SVG
-Streamline converting at least basic images to SVG
-Add support for GIMP / potrace path style export as used in various FPGA-netlist tools SVGs
	<path d="M301 2660 c-1 0 -1 -11 -1 -50 l1 -49 14 -1 c8 0 14 0 14 0 0 0 1 12
	1 26 l1 26 11 0 c10 0 12 0 13 2 1 2 2 2 8 2 6 0 6 0 7 3 0 3 -1 40 -2 41 0 1
	-65 1 -67 0z"/>
-Look over FPGA-netlist-tools some more to see what I can leverage
	In particular FPGA-netlist-tools/masks/hp35-bitmap-tools/masks_to_netlist.py is very short and seems to do a lot
-Support reading source data as masks vs chip
	Self aligned gates may be represented as solid diffusion blob
	How like is it that they don't take the entire area?  For example they only need approx area around the edge
-Only img2js tool is currently maintained, others are more of placeholders
	Deleted, remake if needed
	uvjssim-img2polygon.py
	uvjssim-polygon2js.py
Make output more predictable instead of random hash map lookups


-Merge same layer polygons
	DONE


