That’s a really good question and a really tough one to answer. So, here’s how I’ll sidestep it: 
For me, and for most of you, MPE is the present and since the present is all there ever really is, well…
I feel fortunate to be alive during this time period, mainly every time technology has a ‘win’. I consider MPE one of those wins.
As long as a number of us own MPE enabled controllers, there’s a market for MPE enabled synths. And as long as people buy MPE enabled synths, there’s a market for MPE enabled controllers.
We’re a long way from critical mass, but we’re also a long way from failure.
We can ensure its bright future by producing great music that inspires people to ask after the tools that we’re using.
That’s pretty much it.
I understood the question as something like “Is MPE a passing cloud, something that will vanish in time?” Mainly because I know nothing and don’t care about Quadraphonic sound (which was, I think, the point of the comparison).
But yes, as @c0nsilience says, MPE is the present. And for the time being, there are a few people interested and enthusiastic about it.
I came to MPE rather indirectly, because the layout of the Linnstrument is perfect for my needs. The fact that I can modulate the sound by changing the pressure on a given note or apply vibrato by slightly moving my hand (finger), just like I did on my fretless bass, is for me an extra bonus. Actually, having almost zero knowledge about MIDI, it almost surprises me, that this is possible only recently.
Also something that it’s getting clearer for me as I move along in my adventures in electronic music is that we have very different ways of “playing” music. For modular musicians even a keyboard could be unnecessary. Others use sequencers and/or “paint” notes on a piano roll and add automation lines. Or launch clips and loops that manipulate with sliders and knobs. There are even people that code their music live.
I think it’s great that there are so many approaches to music creation. I like the physical movement of fingers to notes and for that, I think, is a privilege to count with MPE.
Something else is the gentle, patient passion of Roger Linn. I must say that the way he talks about it, his presence on the Internet, in the support forums, even his business model where the instruments are just handcrafted has a big influence in why I decided to go for it.
MPE maybe a passing cloud (I like that analogy) , because its a protocol, and one that exists because of MIDI original motivations … e.g. CV does not have the same ‘issue’ - so Id actually hope that MPE becomes redundant because its kind of consumed by MIDI taking a more open approach.
(e.g. Id love to see pitch as being ‘fractional’, rather than having pitchbend as a separate thing)
on the flip side, I think quad sound is very much ‘back’,
lots of people now have 5.1, 7.1 systems at home, and the ‘tv’ has turned into a media centre, so perhaps we are approaching a time when producing music for multi speaker setups (for general consumption) is more viable.
its something Ive considered, except I don’t actually have space in my studio to add more speakers (and dont use the ones I do have much either 
)
MPE might be a passing thing, but I’m reasonably confident that expressive controllers are not, and MPE seems like a good development in trying to land on a common standard that manufacturers can agree on and work with. I’d like to hope that we see more MPE enabled hardware over the next few years.
I’m really hoping that quad is back. I’ve very much enjoyed quad concerts that I’ve experienced at Superbooth over recent years. I’m sure I can find room for another pair of Genelec 8020s if I could figure out a good way of managing the setup. Right now time is very much the limiting factor for me!
Coming late to the party. Liking the conversation, here. Quite thoughtful, with quite a bit of insight.
And it sounds like we’re on the same page. MPE itself is a bit of a kludge, a temporary solution for a shared problem. It’s unlikely that it’ll become adopted, as-is, by all the incumbents in the music technology industry. But it’s also unlikely that our MPE devices and software will become irrelevant in the future. Yes, MPE is likely to be superseded by MIDI 2.0 at some point. But MIDI 2.0 is likely expanding from MPE, not taking a completely different direction.
It’s been interesting to observe industry attitudes to MPE. A few manufacturers and developers have readily embraced it and there’s a sense that they’re “all in the same boat”.
ROLI’s role is quite unique, IMHO. The company has been championing MPE in many ways and they probably have a specific role to play through JUCE. They do recognize other players in the MPE scene. Just short of showcasing other people’s products in their own marketing material, their communication about MPE is rather inclusive. On the other hand, the association between ROLI and MPE has become strong enough that some people confuse that company’s approach to “5D sounds” (including monophonic ones) with MPE. I’ve recently tested a monophonic softsynth for wind players which has a mode for “MPE”. Since it works with MPE controllers, the devs think that’s quite fitting. 
At the other end of the spectrum is Ableton. The company’s absence from the MPE world is conspicuous. In some ways, Live’s the only major DAW which calls for kludgy ways to add MPE-like capability. Most other DAWs either support MPE directly or matter a lot less for a mainstream audience. Maybe they’re just waiting for MPE to be more established. Or they’re betting on another standard, which doesn’t yet exist. The fact that Ableton (and Native Instruments) have recently joined the MMA might be a sign of something positive happening in expanding MIDI features in their own products to include polyphonic expression. I’m “cautiously pessimistic” about the level of lobbying or mere inertia that can come through such interactions. If anyone can make our MPE tools obsolete the way “quad audio” systems have been, it’s probably Ableton.
(In that context, it’s kind of interesting that Cycling ‘74 does support MPE through Max 8. As far as I can tell, it’s the only part of Ableton which does. Wonder what internal conversations have occurred about enhancing or deprecating MPE through further integrations of Live and Max.)
In the middle of the MPE-championing spectrum are organizations which use MPE as a point of distinction without “betting the farm” on our beloved spec. My favourite example is Bitwig in part because all the softsynths in their flagship DAW support MPE in a hassle-free way. At the same time, MPE isn’t their major selling point to a wide public. In other words, they’ve embraced MPE but they don’t make a huge fuss about it. Part of what excites me about the modular softsynth and effects processor in Bitwig Studio 3 (“The Grid”, as they call it) is that it’ll be a convenient way to build our own MPE synths and effects. Where it’ll be especially useful for me is on a Linux laptop with a touchscreen.
Apple’s Logic Pro X is somewhat similar to Bitwig Studio in the sense of embracing MPE without making it a core feature. But Apple is its own beast and it’s unlikely to ever become a major champion of MPE.
Something I found very interesting is the fact that Duda has recently added MPE to Serum. In my experience, it’s been something of a dominant softsynth, with quite an intense following. Introducing Serum users to MPE can be surprisingly effective in getting more electronic musicians to embrace the standard. Plus, the MPE implementation in Serum is quite appropriate, IMHO. If anyone can inspire softsynth-makers to implement MPE, it’s Xfer Records.
These days, much of my musicking happens on iOS and it’s been fun to get a pretty steady increase in MPE-savvy softsynths (especially as AUv3 plugins). What’s cool there is that it’s pretty much an “open market”, with fewer incumbents.
Some devs are embracing MPE quite actively. For instance, recent releases by Kai Aras have been specifically geared towards MPE, including a plugin for virtual keyboards and controls.
At that game, there’s also moForte (of GeoShred fame) with Jordan Rudess as a strong MPE advocate.
The AudioKit people have yet to release anything MPE-compatible but they’ve been on the record that MPE support is on their roadmap for Digital D1 (and, presumably, Synth One). If an Open Source MPE softsynth comes out on iOS through AudioKit, it could open up quite a few possibilities.
My sense is that, as the iOS music creation expands, there’ll be a broader base of support for MPE. Not that it can guarantee survival for MPE, but people who’ve trained themselves to develop MPE plugins for iOS can probably adapt to whichever standard for Polyphonic Expression comes out in the future.
There are organizations which implement MPE half-heartedly. For instance, I don’t think Initial Audio’s Sektor can really be called “MPE compliant” (at least, it doesn’t behave as expected). In some ways, it’s unfortunate that some people don’t pay enough attention to the details of the spec. But it also means that MPE has become enough of a selling point that these companies find it useful to claim to support MPE. (Andrew Belt’s VCV is an interesting case. Rack will support MPE in the future but it’s not clear to me that the integration is more than checking a box for “works with ROLI controllers”.)
So… My sense is that MPE has quite a bit of future resonance. It’s actually quite likely that MPE will morph into something else, but devices and software which currently support MPE are unlikely to be “left in the cold” when that something else comes out.
I, for one, welcome our MPE overlords.
I’m guessing that this comes down to their basic architecture of the software probably needs some largish effort to incorporate support properly. They will have been focusing on getting the VST3 support done that they’re only just releasing I imagine, which is clearly a much more important milestone for the whole of their user base.
It’s possible that if it really does require very significant work they may even wait until they’re adding MIDI 2.0 support… Who knows!? 
i think adding MPE support to a synth is often ‘relatively’ easy - it’s mainly about internal routing messages to voices.
(similarly low level tools like max/pure data/reaktor its also quite easy to add support, without breaking anything)
but for many daws, Id suspect, it’s a pretty big upheaval - storing automation per note, UI, file structures, routing midi between channels - all whilst retaining ‘backwards compatibility’.
add to this, the ‘relatively’ small user base, its probably hard to justify a really large development resource to the effort required.
what I’d do in their position, is just continually nudge my design/architecture in the right direction, esp. when you go thru one of the (inevitable) upheaval/re-write phases.
unfortunately, we don’t know where ableton are in their dev cycle.
I think 10 was still part of a consolidation phase (as was 9) - it feels really solid now, and also they seem to really have sorted out their development process.
I would not be surprised if 11 was more ambitious, but really impossible to tell from the outside, so I might just be being overly optimistic
MIDI 2.0 might be an incentive to finally incorporate per-note expression support. On the other hand, (afaik, only played with the demo at some point) Ableton managed to ignore Aftertouch which was part of MIDI from the beginning for quite some time (or still?)…
Hope the MIDI 2.0 guys take a close look at MPE+ from the Continuum guys. Imho nice, careful additions that seem to make particularly sense for instruments that can sample and transfer events with high enough frequency that note-on velocity can be replaced with realtime sampled envelopes.
Ableton Live has or did have a weird MIDI implementation that remapped incoming messages to channel 1:
https://forum.ableton.com/viewtopic.php?f=3&t=69095&start=90
So trying to get it to do MPE is at best a kludge:
Correct. And very much still a problem. (It’s especially frustrating for max for live developers.)
This makes it suck a lot less, though:
“MPE is broken by design”
This is from the AudioBus forum at around the post on the 20th April.
I have no arguments as I have no technical knowledge. I just know that I love three dimensional, per note expressiveness and the effort of breaking free of “on and off switches”
I’m sure you guys have a lot to say about this.
MPE has its issues but many of the things raised in that post are misunderstandings of the spec 
to start with the history is all wrong, the basis of per channel/string messages goes back a very long time, well before MPE/expressive controllers - i think originally early midi guitars.
polyphony, as was pointed out, you can still have > 1 note on a channel, you just don’t get independent control - and post note off its irrelevant anyway.
also you only have 15 independent notes, not 16 - as channel 1 is a global channel
also with USB there is no technical reason to not use multiple ports to go beyond 15…
as for introducing more ‘poly messages’ the point of MPE (and its predecessor) was to NOT break exisiting gear by sending it messages it doesn’t understand… thats why the MMA is relunctant to introduce new messages, or re-use stuff - no poly messages would do that… also it’d be short sighted to assume you could just use 3 axis
then its imperative that pitchbend is 16 bit, 8 bit is far too coarse even for 48 semis slide… and a poly version of that would be 4 bytes, and thats not in the midi protocol at all!
so the guys that did the original midi guitar stuff, and later (?) haken with their implementation have done a stellar job, and being compatible with something that was a pretty simplistic protocol really.
but its not perfect, i think the MPE stuff that was added for negociating zones and PB is not very good - and i hope this will become irrelevant with MIDI-CI
also, really its should all be 16 bit resolution (*), or at least 10… which hopefully hires midi will bring.
(*) you can kind of do 14 bit, with 2 cc’s but theres no real consistency or support for this - I know haken have proposed a MPE+ for this, but needs everyone to agree to it, otherwise its ‘just’ another manufacturer variation.
I also would really love to see some input from Haken on improving midi for the initial transient… Leopold makes a really strong argument that this initial phase of the touch is super important, and should be much more detailed that the rest of the touch.
(the irony being conventional midi reduces this to one number - velocity)
I always wondered why there aren’t more synths and tools taking the leap into MPE. Arturia, Sugar Bytes and Ableton are three companies that I know of, that have no interest at all in it.
Listening to these arguments cast a light into why that resistance, even if they are misunderstandings. “Broken by design” is followed by a long list of things you can’t do with MPE.
From my perspective and having no relation to MIDI before this at all I can only think of the things I can do with MPE. In short is just staying on the playing surface, without reaching to any knobs, to accomplish a certain range of expressiveness very similar to what I could do when I played the fretless bass, the flute, the darbouka…
MPE is not perfect… that’s up to you, wise guys, to decide (what is?). That it needs further development? Yes, please, otherwise is stagnation.
I am thrilled with what has already been accomplished with MPE working on a 35 year old serial interface. I hope manufacturers will come together and standardize a robust, high resolution future MIDI spec that will have room to be extended to embrace technologies not yet conceived. The viability of electronic music for the long term is at stake, as musical instruments with limited expressive capabilities will quickly be superseded by those that allow the musical soul to speak directly and emotively. These truly are the days of wonder.
I had a Casio PG-380 guitar in 1988 which did multi-channel midi, or “Mono Mode 4” as it was snappily called then.
Yes, Midi Mode 4… I remember it well. In the late 80’s I configured a SynthAxe sending six separate channels to a Casio VZ10m to ‘bend’ strings by reading the pressure from the key trigger assembly (not the 2 x sets of strings) and to raise the selected ‘string’ up a programmed semitone or a tone. It was the very first instrument I ever found to allow realistic pedal steel-type bends within a chord.
Do you still have that SynthAxe?
I know they were madly expensive at the time I wonder what they would be worth now?
Unfortunately no … I stupidly sold it to buy a Yamaha AE1200 jazz geetar and a second-hand car. I do have five Ztars though 
which kind of makes up for it and much more. I had the full system with floor controller and the desktop controller. (£16K ) I got the system from Mike Dixon, one of the designers, in stages, and started with the controller, and then added the floorboard and then the desktop and cabling when he sourced them and I could afford them.
It was a great instrument, but vastly over-engineered.
The key triggers should have been made like Hammond keys, and effortless to play. Instead, you had to smash your fingers down onto them, and vamping soon left you with a severely aching wrist. But the keytriggers and the diagonally split frets were great, and should have made it to other geetar synths.
But a huge disadvantage was that an instrument with 12 x ‘10’ guage strings on doesn’t work for me or really feel like a guitar at all…
The Ztar does have keytrigs on some models, but they are rubber bars, and not at all like piano or Hammond keys… Right hand strings were also very difficult to play and shredded your fingers if playing without a pick. Apart from the keytriggers, it did nothing really that Jam Origin and a decent guitar can’t do now (admittedly 30 years later). The original prototype was made from a Yamaha CS80 which fell down a liftshaft.!