I mean that both the initial bugs problems (freezing, etc) as well as RAW support will be fairly easily fixable by Apple. The fact that most issues seem to be easily "reproducible" means that these will also be easy for the software team to debug for our 1.01 upgrade. The RAW problems might take a bit longer, but the pessimism in the ars review ("it's too much of a task for Apple to implement successfully") is just nonsense!
In the short term, my goal is to learn Aperture by developing a new DB of images for a long term project - (one where I won't be needing output immediately). I already have about 20,000 images that I never started developing in Cumulus (my current media management software), so I'll begin importing those in smaller batches and experimenting with tagging them, using the stacks feature, implementing my BU strategies, etc.
I will be avoiding any major exports for a bit. If I need to get an image, I can easily import it directly from PS (based on my naming system). I'm pretty confident that by the time I've got this program figured out, most of the major complaints will have been addressed.
It should be noted that even the Aperture literature suggested an incremental approach as you learn how to use the new work flow. If one imports 100,000 images the first day and starts screwing around with them, a lot of time will be wasted. Better to start small and learn to do it right!