I like to hear people's opinions. With a poll people just vote and many don't voice their opinions.

honestly a Pole with: 5-8 year options, would be fun to see.
I think the 6-7 year option would win out.

At the earliest I dont think a new consol by sony or ms, is due before 2019.

Maybe around 2017 if Nintendo wants to start the new generation, and followed by Sony and MS in 2018-2019.

Why do they have to scrap it and start over? They could use the Xbox One as basis and build on that an enhanced Xbox One called "Xbox Two" or "new Xbox One" (faster but compatible x86-APU, 12 GB RAM, bigger eSRAM).

It would play the same games as the old Xbox One, but in 1080p & solid 60 fps. Every XBO game would still run on the old model (in lower fidelity), so owners of that wouldn't be forced to upgrade. And Microsoft could offer fair trade-in offers for the ones who want to upgrade.

Microsoft also manages to "dump" millions of dollars into R&D in a faster Surface Pro every year, so an enhanced & compatible home console would be theoretically possible.

There are so many things wrong with your first statement in how you think things work, that I don't even know where to begin.  These are gaming consoles, not your home PC.  MIght I suggest investing in one if you are so concerned about hitting 1080p and a solid 60fps.  

The same consumer base for the XB1 and the PS4 would be the same consumer base for the XB1, PS4, "XB2", and the PS5, but spread out over more pieces of hardware.  So now you are asking game developers to spend more time researching into how to effectively create great games for these new consoles AND create them on the XB1 and the PS4.  The game developers would have to have a bigger development team without gaining any additional profit. The consumers would also have to spend more money on another piece of hardware for slightly upgraded graphics (assuming these same consumers already purchased a PS4/XB1).  That would also hurt the company as well because now your focus is no longer centralized. Have you ever wondered why this hasn't happened in the past? 

Gaming consoles are unified amongst gamers over a span of a decade because they have the initial vision of being future proof.  Something that cannot be said about a Surface Pro. 

Not quite. iPhone processors average 4 watts and deliver around 75 gflops on 20 nm, so on 5nm it should deliver some 1.2 tflops?, assuming a completely linear, optmistic scaling on power draw. So still slightly less than the XOne. Tablets will have much more luck though.

I'm not funny, I know.

OT - it probably depends more on GlobalFoundries / TSMC then Microsoft or Sony. If we're stuck on 14nm Finfets for 4 years like 28 nm, they'll wait for the first 10nm revision. If revisions come faster this time around, and their gains are smaller, they'll wait for 7 nm.

The slightly enhanced "new Xbox One" with faster APU, 12 GB RAM and bigger eSRAM was only an example. Of course the next Xbox and Playstation can be a much bigger improvements compared to XBO + PS4 and could allow new games which aren't possible on XBO + PS4.

My point was, that Microsoft and Sony won't have to scrap their work and won't have to start over, but they can build their development on their current consoles... if they stick to the x86-achitecture (and why wouldn't they?). Backwards compatibility would be very easy if they don't change the architecture again, probably with enhanced graphics (similar to a PC with upgraded CPU and GPU).

And I think you overestimate the effort for development teams to decide for additional settings for a faster Xbox or Playstation... every multi-platform game that is also released on PC has to have a scalable engine anyway and XBO + PS4 are closer to PCs than ever before. That it didn't happen before doesn't automatically mean that it would be impossible or stupid in the future.

Handhelds also had the initial vision of being future proof, now we have the slightly enhanced new 3DS with better textures, controls and loading times in some games compared to the old 3DS.

Arnt there people saying the limits of whats possible right now is somewhere between 7nm and 10nm?

We might never reach 5nm..... long before then they might give up, or move to a new type of chip.

(light instead of electricity, moveing tiny light beams around doesnt generate heat the same way electricity does, so maybe they just shoot up clock rates, on some gigantic 40nm+ "light" based chip)

Basically even by 2024 there probably wont be a top end smart phone, near the power of the Xbox one, unless something new comes along and drastically changes things.

Good grief I hope not.
Last gen was waaaaay too long.
Gens should last 5-6 years, no more.