I think we're nearing a critical point with digital cameras, and especially DSLRs. It's difficult to imagine a small sensor compact camera better than the Ricoh GX-100 or the Leica D-Lux 3 (though it's not difficult to imagine a better compact with a bigger sensor). [Several people misinterpreted this remark as my saying you couldn't make a better compact camera. Sure you can. Both these cameras are fine designs except for one thing: their sensor and the resulting image quality.] It's also difficult to imagine FX sensor DSLRs better than the Canon 1DsIII and Nikon D3. In the DX/APS DSLRs, we've got the Sony A700, Nikon D300, and Canon 40D, with very competitive consumer models living under those (and those products are all due for a refresh shortly, which should bring them even closer in competence to their higher-end sibblings). New DSLR products are thus starting to have marginal increases in ability (though I don't dismiss those--every push forward in ISO or focus ability, etc., is useful to someone).
My assessment is that in 2008 we won't be seeing many curves thrown at the DSLR market. Instead, in the near future we'll see more feature consolidation, some modest bump in megapixels, and perhaps marginal image quality increases. Indeed, to some degree we already see those things. Some might argue that the D300 is simply bringing the D200 up to the current state-of-the-art levels of features, megapixels, and image quality. Look at how many things didn't change on the D300, D3, 40D, 1DIII, and 1DIIIs from the previous models, and how many things that did change are simply trying to match new features introduced elsewhere (lithium batteries, dust reduction systems, higher-end AF, live view, etc.). Nikon and Canon really only missed sensor stabilization in the feature consolidation on the latest round of DSLRs. Both pressed image quality aspects about as high as you could expect given the physical limits being pressed against with Bayer systems.
So what is the next market disruption going to be for DSLRs? Most likely the move beyond Bayer. This allows for dropping the anti-aliasing filter, amongst other things. [Several emailers questioned that remark. I'll stand by it.] Done right, a post-Bayer DSLR could up the image quality ante a substantial amount if noise can be held under control and dynamic range not compromised (the Nikon approach is promising in this respect, the Foveon approach not so much; but the Nikon approach is currently unmanufacturable). I don't think we'll see that disruption next year, or even in 2009. Post-Bayer (other than Sigma's Foveon-based cameras) feels more like a 2010+ introduction to me.
So we're likely going from an "I'll upgrade my DSLR every cycle" to "I'll upgrade my DSLR every other cycle" kind of market. You won't be getting big bangs for the upgrade buck if you iterate every cycle. Indeed, my initial assessment of the D300 is that the primary move forward for it is the AF system. If you're satisfied with the AF system of your D200, then I'm not sure that the D300 represents the huge leap everyone suggests. Sure, it looks like high ISO has improved by perhaps a stop and you get another 2 megapixels, but those are small improvements that not everyone values the same.
The compact market still has two potential curves left in the near term, the biggest of which will be the introduction of larger sensor models. The other would be a move towards dedicated AF systems. All else equal, just adding those two things to any existing compact camera would make it stand out from the competition, at least for the serious users.