I feel a bit differently about the in-camera styles, tbh. I understand the point you're making, about the limitation forcing you to consider other elements more carefully (composition, lighting, careful control of exposure, etc) at the time of shooting, but on the other hand what you end up with is really just one possible rendering of the RAW data, and it's a rendering preset that Canon chose, not you. There is nothing wrong with imposing these limits on yourself, but I'd personally want to shoot RAW + JPEG in this case, so that I would have all the benefit of how the limitations affected the composition of the shot, but none of the drawbacks of possibly getting stuck with a photo that should be a keeper, but can't be saved due to technical limitations. Memory cards are cheap, and you can always through data away.
Again, all of our cameras can shoot RAW + JPEG, and unless you need the highest possible framerates and buffer capacity for an action sequence, there is little downside to having the RAW data available. You can have SOOC shots for your clients, but still have the digital "negatives" available if they ask for changes. Post-process burden is all relative and totally in your control. You can tweak each shot for days, or you can just batch render the RAW to JPGs with a few clicks. Plus, pulling a shot by a stop or two, or correcting white balance is less of a burden than reshooting.
These are quite salient points of course, absplastic. And toward your point, the saving of "digital negatives" is enhanced with a camera like the 7D2, because you could record JPEG on one media (which I subsequently copy from for the client), and RAW on the other media. In fact, I will sometimes do this for a black and white Picture Style.
Creatively though, I find there is just never time for the post-work mandated by such a workflow. There is something peculiar gained making the choice of "digital film" when the model is live on the set, and the lighting setup is fluid. This intimate coupling of elements is something that having the RAW later will be unable to recreate. If you "see" tonal relations from SOOC work live in situ, note where the shadows fall, and then you relight, and change your setup in response -- having the RAW later does not expose to you what could have been a different, or more favorable setup. It's too late to move the light after the shoot. That's a rendering you will not have had, RAW cannot save you from a choice you did not make at the time you could. This phenomena is not about how tones end up in a 14-bit 2D spatial digital rendering, it's about what could you have done with everything present in the 3D physical space before the foto was taken.
The "limitation" of making an experimental choice of "digital film" is a kind of freedom from bondage. Rather than presume everything can be "fixed" in post, you are instead forced to "get it right" in the camera at the shoot. The maximal exploitation of this freedom is the inducement to be unafraid to move the light to compose a creative foto. Instead of using a zoom lens, it's much like using a prime to force you to move your working distance and angle in order to obtain a potentially novel or beautiful framing. Fixed digital film forces you to examine all of the elements composing a scene, particularly the light, since that is what the "film" responds to.
Let me augment the dialog here by commenting on one of your assertions. I'm not using rendering presets with Canon picture styles, far from that. I don't even think about those few Picture Styles supplied by Canon. Instead there are other, far more creatively-inclined sources available. Among others, there is a very lively cottage industry that creates truly wonderful Picture Styles for Digital Cinema applications. Those Picture Styles are no less applicable to stills work. Some of them in fact offer gorgeous treatment of skin tones right in the camera, a big plus for the glamour-related work that I do. Nothing less than fantastic options, all around. Plus, there is much artistic skill to be gained by having the knowledge to create a custom application-specific Picture Style. In this way, you come to know your camera imaging system very intimately, and can make it do your bidding.
I can see RAW really helping for very high quality results in commercial workflows, where a team of people work together in a pipeline, with lighing assistants, including perhaps a Photoshop master, a professional retoucher, and so on. Jobs where it's not only the final image quality that matters, but rather how quickly final images can be produced. In other words, RAW is a good hedge against modest lighting errors in a case where working quickly on set is important - because "time is money." Here there might be a model or models, they do their thing; and the photographer and her/his team do their thing. Everyone has their separate job in this scenario.
I sure don't have budget for such productions, and more particularly, I don't even want to work that way. My models (belly dancers) are generally Muses -- we always work together, going as far as pre-planning looks before the shoot. So, the model takes a role in choosing a suite of options on tonal relations before we even start the project. And when we're in the studio, we work interactively in our development of the work. This is a kind of pre-work that augments the work in the studio, so that with intentional choices in Picture Styles, we get the best results without requiring a post-work project cycle. A much smaller team. Everyone is an artist interacting with another artist in this scenario.
The last point I wanted to make about Picture Styles is that the state space of all of the possibilities with RAW is so enormous, it is highly improbable that every possibilty that could be will present itself in post-work. Picture Styles could be viewed as a form of guidepost: something to show you the way, and from there you respond. This use of RAW appears to affine your views absplastic: choice after the fact.
My experience though is that making choices given a myriad of options in post with RAW after the fact is a very different than intentionalizing a result before it happens; perhaps somewhat analogously with how it was with film. Learning to move the light in the studio to improve the "read" of your "film" is very different than tonal adjustment in post, never having confronted the question of where could the lights have been.
