the Eris has a micro SD card ( not full size SD card ) and you can't get to it without removing the back cover of the phone.
The edges of the iPad are not 0.5" thick since the device is tapered. There is also no "cover" to take off.
The μSDHC card in the Eris is the main memory, not an option slot for uploading, and the screen is similar in size to many digital camera screens. The hidden slot makes sense for the Eris.
Battery life , relatively large screen , how it feels in your hand are all higher on the design priority list than SD Card slot.
All it would take would be 1mm less taper, and you could fit an SDXC slot *and* have more room for a bigger battery!
"Thin" trumps usability again. And with less taper, the case could be a bit stronger - I would how much of a fall, or merely pressure, will shatter the screen. (If you really believe that the lack of an SD slot is a design decision, and not part of the walled garden.)
Pure NAND is most definitely better than some most likely USB controlled SD port.
I would prefer the inbuilt storage option IMO especially if its going to be faster. I've never needed more than 16GB for anything.
If you don't need more than 16 GB - buy the 16 GB and forgo the SD card.

Let someone who wants more buy the 16 GB and add a 64 GB SD card.
Agreed, RAW is the way to go. But what software will handle it? From what I've seen iPhoto turns my RAW files into teeny little useless poster size files.
Adobe Lightroom for iPad?
I would guess that many photographers would only use the Ipad for "proofing", not manipulating photos. It would just be a larger screen for checking quality - but easier to carry than a full laptop.
It wouldn't need a full "Lightroom" editing capability, just basic developing tools and pan and zoom. Add the ability to upload, and you could share proofs with the office or client.
What I'm getting at is that even little parts which cost less than a buck become a big deal when you plan on manufacturing 10's of millions of units. ... What parts/components could you afford to buy 10 million units of without it zeroing out your bank balance?
I don't think that analysis makes sense - because the component cost is fixed per unit.
The real business decision is whether you want 40% profit margin on the device, or whether 39.91% profit margin is good enough. Add to the decision the possibility that making the device more useful will improve sales, so that even with the smaller margin you'll make more money.