Version 1.00 is here - survives VERY HUGE projects!

Instead of reading all dependencies up-front (which kills memory in our project), there's now a pipeline in place so that only the dependencies of one file are in memory. This requires double reading if graphing is done, but it is now possible to read arbitrarily large projects (all of .Net, for example ;-) ) and is still faster, as memory comsumption is lower.