The recent offerings from Canon and Nikon seem at first glance very impressive and very exciting but I find on reflection that there are really no improvements significant enough to cause me to rush out and place a pre-order for any of these new cameras, except perhaps the Nikon 300D. I should qualify that. No significant improvement for my style of shooting which is not at least partially offset by a serious disadvantage that causes some hesitation. I like the 300D but have no Nikkor lenses.
I have a wide format printer, the Epson 7600, so extra resolution will always be appreciated and the 1Ds3 would certainly provide that with its with 21mp since my current DSLR is the 5D. I have no doubt that a leap from 12.8mp to 21mp would be significant and noticeble on a 23"x34" print, though I doubt that current owners of the 1Ds2 would find any significant resolution improvement. I expect image comparisons between these two models (the 1Ds2 & 3) will centre around subtleties of tonality rather than sheer resolving power.
So what's holding me back? The price? Only partly. Give me exacly what I want and I'm prepared to pay for it. It's the extra weight, compared with the 5D, which is of major concern, plus the fact that high ISO performance does not appear to have been increased in the 1Ds3. Of course we don't really know that at this stage. It might well be the case that 1Ds3 images really are much cleaner and sharper at ISO 1600 and ISO 3200. Perhaps Canon, in order to maintain higher standards than its competitors, have resisted the 'bells & whistles' approach of offering an ISO 6400 or 12800 which is really not usable.
As Michael mentioned in his "Buying a Watch' analogy, we each have to decide for ourselves which features are useful. New features which are 'cool' might be great for advertising purposes and for those who simply like to collect cameras, but for those of us who consider cameras merely as tools, there seems little point in paying for features which one might never use. It's of little concern to me whether or not the 1Ds3 lends itself well to a tethered situation because I don't shoot with camera tethered to a computer, but obviously some people do and that is clearly a legitimate concern for them.
However, I do use high ISOs so any improvement in that department is most welcome. I generally don't shoot in the open when it's raining so whilst waterproofing seals are nice as extra security, they are not a 'must have'. The high frame rate of 5 per second in conjunction with +/- 3 stops for exposure compensation would clearly be useful for HDR images or for reducing noise by stacking images in CS3E. The faster frame rate would also lend itself better to hand-held shots. Greater accuracy when using the auto-align functions of CS3 and CS3E.
When I hike in places like Nepal where one never knows what interesting scene, person, event etc is just round the corner, I recommend carrying 2 cameras around your neck (or somewhere else as long as the cameras are instantly accessible). A 20D with 24-105 zoom plus 5D with 15-30 zoom give me a combined effective focal length of 15mm to 170mm. An upgrade from the 8mp 20D to the 10mp Canon 40D doesn't seem justified. If Canon had delivered a cropped format upgrade with the specs of the Nikon 300D, I'd go for it.
In fact, from what I've read so far, the 4 features of the 1Ds3 that attract me most are (1) higher pixel count, (2) greater exposure compensation (+/- 3 stops), (3) faster frame rate, (4) auto-focussing at a minimum aperture of f8 which I believe is a feature of all '1' series cameras, and better, faster focussing in general. I'd like to use the 1.4x extender with my 100-400 IS zoom more often.
Pack these 4 features in a body of similar weight to the 5D and I'd have little hesitation. Give me increased high ISO performance in addition and keep the price similar to that of the 5D, and I'd have no hesitation whatsoever.
Looks like I might have to wait for the successor to the 5D, but it worries me that all the rumours so far suggest that the 5D successor will be no more than 16mp. I'm not really interested in small percentage increases in pixel count, like from 8 to 10, or 12.8 to 16.7 unless that's accompanied by some other major and really useful improvement beyond a few bells and whistles.
Right at the moment, the Nikon 300D is looking like a great competitor to the Canon 5D in terms of everything except wide angle coverage. In fact, if I were buying my first DSLR and I wasn't already sort of locked into a Canon lens system, I'd probably choose the 300D.
But who knows what features the successor to the 5D will boast? Is it too much to expect it will seriously challenge the 1Ds3?