I would like to ask everyone who is arguing that a 2048x1536 display is a "done deal" to consider trying to design such a product at current iPad price points. Given the competition shown at CES the next iPad will need a dual-core ARM9 and a significantly improved GPU. Add into that increased memory (because I really can't see the iPad remaining at 256MB when some of the tablets/phones at CES were running with 1GB) and the FaceTime camera that we pretty much know will be included and I think you've already got a product that will be pushing the limits of the current price structures.
Personally, I'd like to see a doubling of the included flash memory since the entry-level 16GB is really inadequate and even the current top-end 64GB is a little restrictive for those who want the full-media experience on a large-screen mobile device (e.g., I just purchased the Life magazine guide to digital photography which occupied 0.5GB of space -- for a single multimedia "book" -- then consider that one HD movie runs almost 4GB). Of course, going to 32GB, 64GB, and 128GB models would be a little pricey so that too may be out of the question (the other "too" being the Retina display, which I would argue isn't price/benefit competitive in the 2011 tablet space).
Let's start with the Motorola Xoom specs since that was pretty much the tablet of choice at CES:
Tegra 2 CPU/GPU (1GHz, dual-core ARM9 with NVIDIA GPU that appears to significantly out perform today's iPhone/iPad SGX535 graphics processor)
1GB DDR2 RAM (standard)
32GB flash storage (standard) with additional storage via built-in SD card slot
10.1-inch widescreen 1280×800 display
Front and rear facing cameras
Built-in gyroscope, barometer, e-compass, and accelerometer
micro USB2 port, HDMI output, and a few other extras
Now consider that Apple very seldom offers really cutting-edge hardware (if you exclude the form factor and industrial design). However, they do offer well designed products that tend to hit the market sweet spots on performance, features, and usability. Of course, they also charge somewhat elevated prices for any given hardware spec, something that should not be overlooked in any proposed redesign on the iPad. Next consider that the iPad needs to be a high-volume product, one or two million units shipped next quarter won't be acceptable and here we're getting into the very definition of what makes cutting-edge hardware so rare -- you can't produce a lot of it because the technology is new and supplies are going to be limited.
As far as using the iPhone 4 as a model for the iPad's evolution, let's start with the admission that the iPhone 4 is pretty much a repackaged iPhone 3GS with the addition of the Retina display and FaceTime camera (yes, I know, it's got the Apple A4 processor but that's just a tweaked version of the ARM8 that is in the 3GS -- along with double the DRAM). Is Apple going to do the same with the iPad? You might say they could -- add the Retina display, FaceTime, and double the DRAM and presto! -- the new iPad. But, frankly, this time around Apple needs to do more than that and besides with the iPhone 4 that is just about all they could have done (since dual-core ARM and faster graphics were largely unavailable or impractical when the iPhone 4 was introduced).
A further problem with the iPad-as-iPhone approach is that this largely ignores the competition's Tegra 2 CPU/GPU and it also tends to minimize the difficulty in producing a 2048x1536 9.7" LCD that could match the quality of the IPS display used in the current iPad. Making the jump to a 9.7" display at the so-called retina resolution (or near to that) is a lot more difficult than doing the same on the iPhone's 3.5" display (as someone said earlier, it's not simply the case of "gluing" together four or more iPhone screens to make one for the iPad). One of the reasons for this is that the manufacturing yield of LCDs is strongly influenced by the total square area of the display -- simply put, a larger display (by area) is more difficult to produce even if both products are at the same pixel density (pixels-per-inch).
Just try to compare the price on a 35" LCD HDTV to one at 97 inches (if you could find the latter, which I don't think you can) and note that a 97 inch, 1080p HDTV has a much lower pixel density than does the cheaper 35" set (I'm not trying to claim that the manufacturing problems and market dynamics are the same for HDTVs as they are for tablets, but this comparison should give you a hint as to why larger -- in area -- is not necessarily easy).
Lastly, let's look at the the Motorola Xoom's 10.1" 1280x800 display. If you do the math and assuming square pixels that works out to be a 149 pixels-per-inch (ppi) display. The current 9.7" iPad has a 132 ppi display which is a difference of 13% (in favor of the Xoom). Would that be visible? Perhaps with careful examination, but maybe not that significant given other factors (such as brightness, contrast, and viewing angles).
So here's the challenge, design the retina-display iPad and make the determination of what needs to be "cut" and what must stay. Realistically, you can't be equal to and/or better in every major hardware category than Apple's competition -- at least not given Apple's historically high profit margins. The baseline might be Apple's current iPad hardware with the addition of just the rumored 2048x1536 display (not a likely scenario, but you've got to start somewhere).