	-- More debug (CRUSH.debug) 
		--> give traces for all exceptions
		--> Remove ? from NaN checks?
	-- Merge util with xcrush and update to it...
	-- Common TextFileParser (and TextTableParser?)...
	-- RMI (client/server) mode...
		-- InfoMessage, WarningMessage, ErrorMessage (with stubs)
		-- ScanSet (for fixed threads)
		-- Node/Job Manager
		-- node.readScan(), or node.addScan()
		-- Service discovery via UDP
		-- node.merge(SourceModel)
	-- merged steps?
		-- decorrelate/gains
		-- weighting/time-weighting
		-- despikes
		-- filters
	-- parallelization
		-- drifts
		-- decorrelate
		-- gains
		-- weighting (all)
		-- despike
		-- filter
		-- reading/validation
	-- Orientable corrections (e.g. single scan rows...)
		-- Multimode analysis
		   (choose the highest mode with a certain power
		   fraction of the zero mode)
		-- mean rotation span from weighted center
	.. Documentation.
	-- Shadow reductions
		-- Shadow data, or shadow scans
		-- add shadow sources.
		-- decorrelation
		-- gain changes
		-- drifts
		-- fourier filtering
		-- whitening
		-- shadow source
		   (processed just like real source)
		   (but no weighting...)
		   (weights/flags from real source...)
		-- write shadow data
	.. tcsh does not like [], add note to README
	-- regrid flux (slightly high) noise (way too low).
	.. pointing support
	.. logging support
	-- highpass filtering to hp signals too...
	.. Chopped reductions
		-- Photometric mode. Think this one through...
	.. Update 'extended' mode configuration
	.. Reorganize options keys (hierarchical)
	.. Optimize reductions...
	.. Documentation
		.. What's new (minicrush and crush versions)


