Product marketing doesn't drive development...product management and engineering design does.
Since Lightroom uses a raw processing pipeline for it's processing, the "concept" of a layer based editing is foreign to such a pipeline. While there may be ways to incorporate multiple image editing in such a pipeline, it won't be based on a layers paradigm as we know it in Photoshop. It would be more a mask based composite more like what's in the Adjustment Brush now.
The power plays do occur...but while Mark Hamburg left Adobe for MSFT, he left MSFT and returned to Adobe. He has great influence and vision (he started Lightroom/Shadowland) and made a decision early on that LR would be a workflow based raw processing DAM application. Dwelling for great periods of time on single images (or composites) is simply not what Lightroom was designed for. It's designed for working through many images...the way LR has been developed was intentionally very divergent from Photoshop.
I'm not saying Lightroom might not see a compositing module but it won't look anything like Photoshop's layers. Panos and HDR are far more likely than compositing.
What drives marketing, development, and engineering is at best a circular chicken/egg thing. What can be marketed, and can be developed, will be marketed if the numbers are favorable. Sometimes engineering produces that which can be directly marketed, or developed and marketed.. so many ways this can go. But with a defined market and a 'developing' product the marketing guys are very much into the decision making process. There are very few engineers calling the shots these days. But none of that really matters, I think we can agree that if there is a huge call for LR to have layers, or layer like capabilities, and it will increase their market share/profits, then we'll have layers or layer like capabilities in LR.
Is layers at the raw level an engineering impossibility or an engineering problem yet to be solved? And when it's solved will it work like layers in PS or will it just provide the same capabilities achieved differently? I don't know enough to speculate, but from a business standpoint I'd guess when they do solve the issue, and it looks like the plug-in in the OP has went a ways to doing just that.. they'll call it layers and people will accept it. For no other reason than because people know what layers are, what they do, and how they want to use them. It's solid marketing. UNLESS the new way offers significant benefits/advantages that warrants new terminology, which will result in more market share/profits.
Your knowledge of LR's development is interesting. It helps us understand why we've seen some features and not others, and like you said, what we might see first in the future.
Yet, I've observed a big call for layers in LR and I can't see the ignoring it forever. We've all started projects with specific goals and intents, only to see them refined and developed along the way. There are many examples.
Personally I use both.. and C1pro and more. I'd rather see them put more effort into database speed and efficiency, make it more stable/reliable, and take better advantage of newer hardware (more cores, RAM, SSD technology, hybrid tech, etc) to make our existing experience faster and more satisfying. I import images into PS all the time, almost as a matter of course, and just having layers, or HDR, or Pano's.. would not change my workflow. I'd still do these things in PS. So far it seems that with actual imagine editing tools, most everything in PS works better for me. I'd like to see CA handled much better though.. on par with C1 would be nice.
It would be soooo interesting to be that fly on the way and watch how these decisions are made.. we can all guess how such decisions are made in other companies.. but I want to know about this one..