Chuck has well explained this, because he understands the side effects of any kind of sandwich compression, whether "minus"compression development or scanning tweaks afterwards. Numerous films are capable of bagging extreme ranges of luminance when specially developed for sake of that. But there is always a penalty. The proof is in the pudding, the print itself. There's a penalty to the intermediate gradation and overall microtonality itself. Digital printers without a strong prior background in darkroom printing often don't duly appreciate that distinction. Hence there are all kinds of histogram and other arguments that aren't entirely relevant. My own eyes tell me something else, and even my own light meter.
And without trying to be dismissive of your own approach, Steven - that scene you posted looks like a garden-variety luminance challenge to me. Try taking a shot of gleaming glacial ice in direct sunlight adjacent to deeply pitted black volcanic rock in the same neg, and try to retain in print the most intricate specular sparkle and reflections in distinction form the overall whiteness, and then almost tactile tonality all the way down to the distinction between black rock in shade and the deepest pits and cracks within that, and then tell me how to achieve that on HP5. I've successfully done it many times, but never once with HP5. And I know a lot of sophisticated printing tricks. Scanning won't make any difference in that respect.
A full scale 12 stop range print on HP5 simply won't have the same convincing sparkle and intricate depths of shadow as if the same shot were taken and masterfully printed from a more straight line film like TMax 400 or ole Bergger 200.
It can't. You're skating on extremely thin ice way down on the toe, where there is little substance to actually reproduce. But the distinction itself is even hard to render in inkjet, being a less nuanced medium, so people claim all kinds of things without a valid comparison. It's the QUALITY of what's in the nether portions of the exposure scale, not just their mere existence, which counts.
And then there's the case of certain UV media like carbon (which Vaughn does) or Pt/Pd printing, which can handle longer scales than silver gelatin projection papers, and indeed generally needs "thicker" negatives. But even in that case, a longer potential native scale on the straight line portion of the film has significant benefits. Hence Super-XX was prized in its time by UV and Azo contact printers. Proponents of Tri-X would heavily overexpose for sake of contact printing media, yet that habit had the side effect of blowing out the highlights if the same negs were used for regular silver printing.
So one should think more in terms of nature of the distribution of full tonality rather than just somehow salvaging its nether portions. If it can be done via scanning and digital curve reconstruction, it was done by various masking protocols already many times before, perhaps not as easily, but just as effectively. You can only retrieve something already there, and only in relation to its actual quality. And having worked with HP5 a lot myself, I've learned its realistic limits. I won't go into my own broader toolbox of pyro developers, masking strategies, VC split printing etc. Those things can certainly help, but can't manufacture tonality that isn't decently caught on the neg to begin with.
Again, I'm not implying there is anything wrong with your choice of film. HP5 has a number of wonderful characteristics, especially when pyro developed. But there are good reasons why other film choices exist too.