Heh, I'd say the main problem is that mobile devices...don't have buttons and analogues. DS and PSP had buttons and analogues, so traditional games had possibility of being transported without losing their roots too much. This doesn't happen on mobile platforms.
I would point to the few (often really awful) instances of developers trying to make games fit for dual analogs on the DS and the many (often problematic) attempts to put dual analog games on the PSP.
Eventually developers started building games and mechanics that made this make more sense.
And yeah, I know controller adapters will come, but I just can't see the iPhone audience so keen on bringing them everywhere, and for playing 5-10 minutes. Yeah, I feel the current mobile audience, especially after how it has been cultivated, is not that ready for the transation. I have many doubts, both mechanics-wise and audience-wise.
A lot of the PSP games that were rebuilt to a short mission to mirror structure the ideals of portable game often didn't work unless a lot of care was taken.
Eventually developers either figured out a good way to do it or realized that it was okay to target a smaller audience that is interested in playing long games on a portable device, and just accommodating them with systems like save anywhere instead of the 1+ hour apart save points of the PS2 era.
However, I still maintain next gen Nintendo handheld console must have still two screens, but this time both touch, equal in size and not separated; so, the console fully opened becomes a tablet. This + Web Framework + Unity = many, many iOS / Android games that can be very easily ported on the platform without almost any hurdle at all (if they want, just implementing optional buttons / analogs control)...but without the whole gigantic amount of pure crap that pollutes both iOS and Android stores.
I think taking what was unique about the DS (primary screen touchscreen + buttons) and turning it into a PSP with a second screen for map and inventory utilities probably wasn't the best idea for maintaining their identity.
I feel making the internal hardware based on the same parts almost all smartphones and the Vita uses will help them notably in getting ports because everyone can use the exact same technology they do on other platforms, and it's also incredibly cheap, powerful, and power efficient, which are all things that would work well for Nintendo.
As for the actual design of the handheld, I'm a bit more open, as long as it's ultimately designed to offer a quality experience designed to accommodate many forms of modern gaming, has a strong digital connection, variable software pricing, and an attractive hardware price point.
And being still on the matter of the next Nintendo handheld, looking at how Vita can get day and date multi releases of quite important / mid-sized PS3 titles too, I think it needs to have the right tools for PS3/PS4 ports, as well as being quite powerful. Not too much, but enough to allowing even more home-like experiences for the Japanese audience, or even more. A Holiday 2016 release should allow to have things like a 4GB of LPDDR4 RAM, with 70GB/s as highest value possible with still a 199.99 / 19,900 Yen price in mind. I'm so much convinced power is important, as well as the entering price, that...I'll say I'm even ready to see them throw away the stereostopic 3D. And I'm saying this as one big estimator of 3DS' 3D, but if that can help the platform in being powerful and not expensive, then... ;_;
Yes, using modern mobile parts would get them very far down this road, simply because almost everyone on earth (including tons of Japanese third parties and even Sony's platform that's receiving lots of ports right now) makes games for that technology, and thus would be very familiar and happy working with it.