So here's my take on the situation:
1ds3 - ff, highest mp, decent fps, best af, best high iso, built like a tank w grip, f8 focussing.
1d4 - apsh, medium mp, highest fps, best af, best high iso, built like a tank w grip, f8 focussing.
5d2 - ff, highest mp, medium-low fps, decent af, best high iso, (weathersealed?), no grip, f5 focussing.
7d - apsc, medium-high mp, medium-high fps, medium-high af, decent high iso, weathersealed, no grip, f5 focussing
60d - apsc, medium-high mp, medium-low fps, medium af, generally reduced features of 7d.
550/600d - reduced features of 60d
1100d - My-First-Dslr.
so what if they merge the 1d & 1ds? and simultaneous with splitting the 5d lines?
i like the idea of:
1d5/1ds4, whatever it's called - ff, highest mp (maybe 30-40?), decent fps in ff mode, highest fps in apsh/apsc crop mode (don't nikons do this?), best af, best high iso, built like a tank w grip, f8 focussing, dual digic, best high iso.
3d - the "pro-video" version of the 5d3. fixed pellicle mirror, full time video af, raw video, (4k video??), maybe pushing the price tag of the 1d line, but well above the 5d3. rest of specs as 5d3 below, (dual digic?), best high iso.
5d3 - ff, high mp (30mp? depending how high the 1d goes), decent af (as current 7d, 19cross, or better, not as good as 1d5), lower fps than 7d, higher than 70d, f5 focussing, no grip, single digic, best high iso. video performance as 600d now, 1080@30fps, digital zoom, slowish video af as now.
7d2 - apsc, medium high mp (25ish?), ~7 fps, 2nd best af (same or better than 5d3), decent high iso (similar to 7d vs 5d2 performance).
70d and lower - cut-down versions of 7d2, with worse af/fps/specs & features accordingly. maybe keep the current 18mp sensor for now to boost sales of 7d2, depending on release order.
with this scenario, release dates? i'm still going for 1ds3-replacement, if it's called 1ds4 or what, being the next thing out. no way they'll bring out a 5d3 with the same sensor as the 1ds4 first. if we see a 5d3 first, then the 1ds4 is going to have much higher mp.
I'm definitely thinking that the most likely scenario will be 7d2/70d/650d all will share the same sensor and progressively reduced features. no way they'll bring out a 70d or 650d with >18mp before a 7d2. and it follows on from there, 7d2 trickles to a later 70d and later 650d (if 650d is before 70d, then it will have 18mp sensor and 70d will have the same as 7d2, or somewhere between 18mp and 7d2 sensor). It makes more economic sense to have the same sensor for lots of models, economies of scale etc.
we're almost at the limits of most lens' resolutions as it is (my 70-300 non-L worked fine on a 350d at release time, on my 7d it's just fuzzy. sold it to my sis and i'm buying the L version on the weekend). so we're not going to see the leaps and bounds of the sensor resolution we have recently.
my hope? add 1 or 2 bits of dynamic range for the 1d line, keep it the same mp as the 3/5d line just HDR. that'll get your pros buying them like they're going out of fashion...