Interesting reading as always. IBM must still be having a higher than anticipated failure rate in the new chip. As has been pointed out, the infrastructure for a quad system is already in place. And if it is not 2 dual cores, then it would probably be out. There does not seem to be any issues getting the current chip in quantities. Besides, if Apple was going to go with quad processors, there would be different types of 'bread crumbs in the forrest'; totally different case, motherboard, bus structure, etc.
One guy said Apple would take a big hit in sales, if they announced the dual core too soon before they could deliver them. I think that is essentially correct. However right now, the Powermac is not what is 'paying the bills'. (Fedex just arrived with Tiger!!). Anyway, I doubt if a 33% drop in PM sales would be noticeable on the bottom-line. But, I do not think you want to try and win converts and have the business headlines reading, "Apple unable to deliver products". Afterall, those very headlines have been seen recently.
Even with all of the well thoughtout strategic scenarios discussed here, I think the best evidence for determining an accurate launch date will come from an IBM insider who can say that their chip failure rate is finally under control.
PM G5 (2) 2.0, 4.5GB, 160G/250G, LaCie 1TB, (2) 23" Apple, ATI 9800 256MB
PB G4 1.5, 17", 2GB, LaCie 500GB, 23" Apple, Superdrive
MacMini 1.43, 80GB, 1GB, Superdrive
HP (2) Xeon 2.6GB, 2GB, (3) 146GB SCSI, (2) Dell 24"