Well, my take on all this...
I don't think it would be wise to have a "paper launch" early and away from the release of the NV30 product line. What NVIDIA is doing is what I would consider the proper strategy when dealing with strong competition- which is to disallow official specs to be downplayed for months prior to the actual hardware being available for performance testing.
This could only lead to lesser sales on product launch as it would give the competition more time to add hesitation to the specs for the hardware. 128 bit vs 256 bit, pipelines, memory speed and core frequency specs are very fuzzy and not always symptomatic of true performance of the hardware in real-world events, yet releasing these kinds of pertinent stats about a product could either be extrapolated to add hesitation about the product's real-world performance or stat-matching can be counter paper-launched by the competitor. (i.e. if a 500mhz core, 500mhz ddrII spec is paper launched, ATI could, almost overnight, announce an R300 ultra and lean on other things like 256 bit vs 128 bit, or quote substantially higher effective bandwidth numbers by using 24:1 buffer compression figures, etc.etc.)
So the smart thing to do is hold back the paper launch as close to the physical product launch so as performance figures in real world can be held in check against speculations to the hardware specs. This isnt an issue in a market where little competition exists, but is a very big issue when a competitor is at the same or higher level of the product cycle. If counter arguments for a product have no performance figures to address, it's much easier to draw an argument away from said product just by using theory and extrapolation that leans towards "worse case" efficiency delivery. Where the real product falls between worse case and best case will be the key to delivery and execution, but without any real world measurable hint, some people can logically assume worse case which leads to buyer hesitation if the worse case margin is very close to a more mature, older product (all in theory with extrapolation only).
So I really wouldnt expect really pinpointed, specific specs at Comdex either, unless the NV30 product line is ready to start hitting the review cycle a few days afterwards. That gap between paper launch and performance figures will make all the difference in the world for initial launch sales.
