Its possible this is an inherently inappropriate goal.
So many non-computer things fail to scale over a mere 10:1 ratio of size that assuming a constant UI is probably inherently the wrong goal rather than the correct goal just as a default guess.
Some aspects are OK. The scribbled notes on the back of the ATM receipt in my wallet are made with the same character glyphs as I see on CAD blueprints. But the UI and technology are overall very different, because the size and job requirements are very different. Ramming one into the shell of another wouldn't be a win.
Look at the ergonomics, if nothing else. I recently have the technological capability to make a wheelbarrow the size of a garden trowel, and make a garden trowel the size of a wheelbarrow. That its recently become possible for me technologically, doesn't mean its good design or a great UI. It is in fact totally inappropriate and non-productive, although it would be an impressive hack. The analogy to the same UI on a phone and on a big screen TV is clear.
On topic of computer UI, something as simple as adding multiple screens / multiple machines on the same desk dramatically changed my workflow some decades ago.
I was an early PowerBook purchaser, because it immediately removed the need to dual boot for my job. I could run all my formerly NetBSD-hosted network sleuthing software and open the Word docs people sent (bonus: auto crossover Ethernet port meant never hunting for crossover cable).
By contrast, Windows had the power-user segment locked up. For a long while, it seemed that there were little utilities traded around forums to do all manner of things that I'd simply open a console to do on a *nix system. As a result, I always felt really unproficient on Windows systems, because I ended up trawling the webbernets for unfamiliar tools that had obnoxiously specific use cases.
Even now, I think Windows 7 had the most power user-friendly DE ever to grace the Earth, and I look forward to KDE 5 since it looks like it might be a worthy replacement.
There were workarounds for most of these of course, if you just took the time to read the manual, which they hadn't really written yet. All you had to do was to edit some xml by hand somewhere and/or run an undocumented command, instead of editing readable plaintext as before. But do you really want to take the time to learn how to do that just to get your xterms running again?
A: Expert programmers who need nothing more than a browser, a text editor, and a bunch of terminals. They can do everything with the keyboard. They barely even touch the mouse unless it's absolutely necessary. They usually don't care what their DE is, as long as it stays out of their way. Judging from the reaction in this thread, a lot of HNers fall into this category.
B: Power users who spend a lot of time in WYSIWYG applications (graphics, video editing, Excel, etc.), multitask a lot, and make heavy use of the mouse. They know some keyboard shortcuts, but the nature of their work makes the mouse, not the keyboard, their primary method of interacting with the computer. Remember the gal/dude who did all the web design for your last project? Or the team who made that promotional video for you? They probably belong in this group.
C: Casual users who use nothing more than a browser and a music app. Maybe they also use a word processor from time to time, you know, to make garage sale flyers and birthday party invitations. Comic Sans is their best friend, Command Prompt is their worst enemy. They tend to use the mouse more than the keyboard, but both can be easily replaced with a touchscreen. You can usually catch them trying to buy computing devices in the local Wal-Mart.
What's interesting about the situation is that A and C share a lot of preferences despite being at opposite ends of the computer literacy scale. Both groups use only a few apps, so it doesn't really matter how those apps are launched. Both groups can live without a mouse for the most part. Neither group spends much time wandering through menus and settings, A because they don't need to, and C because they don't know how to. This leaves B in the awkward position of heavily depending on the traditional desktop metaphor (icons, menus, toolbars, and most of all, mice).
Unfortunately, A are the ones who make all the DEs, and since C are the largest group, they are the default target users of such DEs. It's easy to target both A and C. You just make a simple, touchscreen-friendly DE where all the advanced functions are hidden behind keyboard shortcuts and/or menually editable configuration files. But while the rest of us fantasize about a "universal" DE that can scale seamlessly from 4" phone to 40" TV, B are stuck in the middle, holding tight to their precious mice and 27" monitor, surrounded by DEs that are so utterly unoptimized for their hardware and daily work.
HUD? Seriously? You want me to take my hand off my lovely plastic rodent and type the name of the menu item that I want to use? Once you imagine yourself in B's shoes, all the backlash against Unity, GNOME 3, and Windows 8 makes perfect sense.