You're eyes don't need to fixate on two screens simultaneously for dual screens to offer a vastly superior experience. You don't need more than a map and some quick equip buttons for the dual screen experience to be superior. Literally just having a map and inventory screens, two things you never need you eyes on more than 10% of the time, makes for a 100% better gaming experience. The games are worse when they try and reinvent the wheel with the second screen. Star Fox Zero and Splatoon are a poor use of the second screen. MGS3D and WWHD are perfect implementation. Map. Inventory. Quick equip. Perfect . Sleek. Non intrusive. Superior to single screen.

Don't need a bigger canvas. The XL's screen is plenty big. Just raise the resolution and it'll be fine.

I never said "everything is hunky dory." I said dual screens aren't "outdated to kids." They aren't. Kids aren't playing on smart phones because they think "dual screens are outdated" - they're playing on smart phones because their parents have smart phones and smart phones have one screen because they aren't built for gaming, but for, shocker, phone calls.

I didn't bring up the summer camp to prove "3DS is doing better than smart phones and tablets." I brought it up to show how kids don't think dual screens are out dated. That's made up. Fabricated. Not true.

The other issue with dual screen in a *unified* platform concept is it doesn't make sense to have one variant that has two screens and another (the television setup) that has only one screen. This causes problems making a game that scales to both platforms and unnecessary headaches for the developer.

Its time for a change anyway, the DS/3DS setup is now 12 years old. It's not fresh, people will see another dual screen handheld and think "oh just another one of these with better graphics". Change it up. 

Well, Nintendo could always make a big touchscreen that simulates doubescreen when needed. Something like this:

I think the pricepoint of the handheld NX is still too expensive (170€ should have been perfect at the beginning), but I'd be fine with a system like this.

Several problems with this:

1. The 7900GT sound oversized to me for an portable device right now. I fear an handheld which would use all it's power would make the chip run very hot, and with it the casing of the console. A 7800GT seems much more reasonable to me, and even that one is more than twice as powerful than an iPhone 6 and more powerful than in iPad. Even the very bulky NVidida shield uses a less powerful chip. Probably also the reason why after over half a year (the chip came out in November 2014) there is still no product announced using it.

2. We don't know how well the power scales (or even how much the chip consumes in the first place), the Luigi console may (in fact, WILL due to all the other components, especially a CPU) consume much more than just 10W. There's a reason why overclocking increases the power consumtion exponentially.

3.I don't know if the clustering for the Bowser console even could work in real life.

4. Single Screen would basically kill backwards compatibility, or atleast make it unpracticable unless a very big screen is used, making the handheld rather bulky

5.eDRAM is very expensive to produce, making the prices for the Luigi and Bowser console alone already utopic. eDRAM is the reason why Intel's Processors with Iris Pro Graphics cost twice that of their non-pro counterparts. Case of point with the new socketed Broadwell chips, which cost 500$ upwards despite their CPUs are ot any stronger than other Intel processors for 200$

6. eDRAM also has not nearly enough bandwith for the Bowser console setup, the graphics part would prectically starve off as it couldn't load enough data. Intel's Iris Pro chips come with 128 MiB, and that's barely enough got get the chip over 850 GFLOPS. Either much more eDRAM (256 MiB at least), which would drastically increase costs, or a completly dfferent technology (HBM, HMC or GDDR) would be needed for this.

7. PowerVR only does graphics chipsets. Where is the main CPU?

8. PowerVR doesn't work well with x86 CPUs (just check early Intel iGPU and their atrocious power and buggy rendering), necessiting again another CPU than it's competitors, in this case probably an ARM Chip. Which leads us to:

9. ARM Architecture as a whole does come with some pitfalls. First, the comparatively very low Floating point capacity, which means that such calculations (which compose the bulk of all calculations in games) will either be very slow or transferred to the graphics part, reducing it's capabiltiy in graphics rendering and increasing compexity for programmers (PS3 Cell rings a bell?) (also why most Benches from ARM are Dhrystone, which excludes Floating point calculations, btw). Caches are narrow, clockrates are low (well, the ones of XBO and PS4 are too this gen, so that is less of a point this time around) and their RAM controllers are slow (and single channel on top of that). Apart from the Floating point deficiency most of these points could be corrected by designing their very own ARM chip variation (like Qualcomms did with Krait, for example), but this would again increase cost and the time needed to develop such a chip.

10. Utopic priceranges for the home consoles, even if all the other things would be corrected, you could add around 100$ on each console

11. The direct concurrence between the consoles would be desastrous. Commodore's Amiga died due to a similar setup because everyone developed only for the lowest common denominator to cut costs, which was in their case the Amiga 500, and here the portable console. So even with  the Bowser console, the graphics wouldn't look any better than on the consoles, and often risk looking even worse despite the better hardware. That already happened before to Nintendo: Just look at many third party Wii Ports. If they had also a PS2 port, then the graphics where based on the PS2 hardware, which only had a third of the Wii's power and thus looked even much worse than they could. Even if the resolution would be scaled up, the textures, and models can't and physics most probably won't (not worth the time and expense for third party developers)

So all in all, not a good design idea. At all.

Nintendo is using IBM, not ARM. There is an ARM chip in the Wii U, but it ain't the main processor, it's just used in the main menu and during standby as the archaic IBM processor doesn't know any C-States (power down features when not used extensively)

ARM CPUs only beat other designs in very low power designs, which it ha made it's own niche. Anything which needs more power, including Consoles, are more problematic for ARM Chips as their power consumption scales very badly with increased clock rates. Also the reason why ARM chips hit a wall at around 2Ghz which they struggle to pass for some years now.

ARM hips are cheaper due to lowered complexity and instructions, but it bites back when those features are needed (or could be needed) as suffer from much lower performance in these cases. So the low price is a bit of a tradeoff.

So NX Mario would be 720p handheld and NX Luigi 1080p? For me the brothers would be enough for launch and easily marketable to substitute the 3DS line in 2016 plus adds a home version for that fusion idea you love (android boxes, PSTV). The Bowser update from WiiU would be confusing to release together with these, you could delay the release to 2017 hoping for third party ports, I say ditch it and force them to make versions for the common denominator hardware, the NX Mario.

Haha NO EDRAM!!!  DUH!!!

OP, I think you haven't thought much about people who want Nintendo console that at least has a chance to have proper 3rd party support. Going into 9th gen with 2.4TFLOPS won't make any 3rd party happy.

I dont thik nintendo would make it this confusing

It's not the same for dozens of reasons. Certifications for one. The certification process for making a PC game? There is none. There may be a little bit of one if you want to get your game on steam or gog, but that's a mere pittance compared to a console. With a console, each specification would have to certified individually. You're talking tens of thousands of dollars for each submission, and there are many many submissions per spec. You are talking about adding hundreds of thousands of dollars to the cost of just getting your game on shelves.

Then there are the control schemes. Do you plan on having every single specification to have the exact same methods of controlling the game, with generally the same layout in terms of look and feel? No. No you do not. one of your versions requires touch controls exclusively, it being a tablet and all. That's a whole control scheme that a team has to spend months of man hours working on, and that's just one spec. Do the other handhelds have touch sceens, but not the "bowser". Guess what? Now you have to spend more months of man hours working on those. And no, it's not as simple as taking a portion of the controls from the tablet and using them on your handhelds. The screens are two different sizes, the finger gestures and sensitivity would have to be adjusted and tweaked even if the underlying system is the same.

Then there's the big one. Optimizations. You keep talking about PCs and how they scale. They "scale" by pretty much not optimizing for anything. As long as your GPU understands your instruction sets you're good to go. This older GPU can't handle the instructions fast enough? Just raise those minimum requirements! Totally different scenario than consoles. Do you want to know the power of optimization? Compare The graphics of Resistance: Fall of Man for PS3 to the graphics of The Last of Us for PS3. Both games run on the exact same specifications, yet the Last of Us looks better than Resistance in every way imaginable. How is that? Because the developers of that game learned all of the tricks to get the most out of limited hardware. Perhaps they how best to optimize the order they send instructions to the CPU and GPU so it runs faster. Maybe they learned about all of the limitations of the PS3's VRAM and know that since it can only store 256 MB that its better to run certain calculations that would typically be performed on the GPU on the CPU instead. All of these things? You cannot do if you're "scaling". The size of that VRAM, and the transfer rate would be different for every spec. Every one! An optimization for one spec could run absolutely terribly on the other. You cannot possibly comprehend how big of a deal this is. This is the core reason why console games look better than they should given the specifications, and you're taking that away from developers.

Then, again, is the cost of dev kits. Do you know how much a typical dev kit costs? Over $1500. Do you know how many people work on typical video game? Over 50. Now you're saying that instead of spending $75000 just on dev kits, you're now talking $300,000. And please don't try to argue that the dev kits would probably cost less because they're similar, you're discounting the tremendous amount of cost of Nintendo creating developers tools that would work across all 4 platforms. That is no small feat.

Those are just the major factors. There's also developing for different screen sizes and different resolutions. There's also modifying 3D models and textures to run optimally for each specifcation. Then there's adjustments for sound. Then there's modifying UIs so that they're easily readible when they're 10 feet away and don't look terrible when they're two feet away. And so on... and so on...

At the end of the day it would probably be more expensive to just develop an NX exclusive than it would be to develop a multiplatform game for PC/PS4 and X1. It is not feasible. It simply is not. Third parties will not touch this with a 10 foot pole.