My head keeps taking me back to that the Mini is going to the first with ARM, but with no "rosetta stone" announcement on the horizon or in the rumor mill,
There have been x86 emulators for ARM since the late 1980s, Microsoft already have an up-to-date-ish one for Windows, odds are Rosetta itself was designed to translate anything to anything... although this time around, Intel might get sue-y about anything that emulates their instruction set. I'd be very, very surprised and disappointed if Apple don't already have an ARM-based Mac lash-up, or an iPad running MacOS, chugging away in a tightly-locked room somewhere.
However, I don't think emulation/translation would be so central to an ARMentosh as it was to the 68k/PPC and PPC/Intel transitions. For one thing, its never
that efficient and ARM doesn't offer the head-and-shoulders speed advantage that came with the previous switches, so its never going to be a great solution. The #1 motivation of a switch to ARM would be that it let Apple effectively build their own, bespoke, chipsets rather than wait for Intel to release the right chip.
Secondly, most
modern software is written predominantly in high-level code and using hardware-abstracted OS frameworks for the fancy stuff. In many cases its "just" going to be a case of ticking the ARM box in the IDE, re-compiling then testing and firefighting any glitches (the latter may be nontrivial, but the back axel isn't going to need to come off - maybe about as much work as supporting the next version of MacOS). As for the potential big stumbling blocks, MS already have MS Office on ARM (not for Mac, but one would assume that they've made as much of the code as possible common), Adobe have just announced
full Photoshop CS for iPad (and, hence, ARM) so that's two ticked off. In any case, a lot of "legacy" software that relies on arcane blobs of lovingly-hand-coded-assembler and isn't being enthusiastically supported is about to be wiped out when the next version of MacOS dumps 32-bit support.
NB: I'm not suggesting that an ARMintosh is likely or imminent, but I think its a lot more feasible than some people think, especially given the power of Apple's current A-series processors. Someone probably needs to implement Thunderbolt (which I believe Intel is now licensing, but I doubt supporting ARM is a priority for them!) - also, if it
was to happen I suspect that it would be a WWDC pre-announcement, possibly with prototype hardware available to developers.