Mac users will wait for the new machines to come out, and compare them to what they have, or less expensive G4 models in app performance. SPEC is irrelevant to these people.

Wintel users considering a switch will do so because of integration, user experience, or apps. It will have to be 'fast enough' for whatever they are doing, but box to box comparisons will not play into it. I can see an exception for BLAST users, or comparable limited scope users.

Ahh, now developers, particularly houses that have cross-platform apps, with the mac one being 'the port'. Will they look at SPEC, and say: "Maybe my mac version does not HAVE to be a dog". "Maybe we can leapfrog our wintel only competition by putting some effort into the mac version". I don't know -- only that mucking around with the SPEC fine-tuning would make me as a developer skeptical that the base hardware potential to leapfrog wintel is true.

I think Apple needs to have a marquee of apps, wherein they go to a developer and offer assistance in making that app soar above wintel. Ten apps with the mindshare of Mathematica that do this will change the world.

Apple needs to appeal to the new, unknown developer who has a good idea, and then makes a choice between coding for the dominant platform and fighting inertia from established players, or coding for the G5 so that the good idea, implemented well, matched with the best hardware stands out. Note, this is only true if the base G5 hardware potential is really greater than Intel's..