Upload & Sell: Off
I approach exposure from the baseline of "normal" being what I see by eye in a scene. But since my eyes adapt dynamically as I scan I see a greater range than the camera can capture most of the time. So post-processing files usually involves "normalizing" them to some degree.
At capture I always try to record detail in solid white highlights. The only 255.255.255 values being specular reflections. The pitfall exposing for highlights like that when scene range exceeds sensor range is every tone in the scene darker than white will be rendered darker than perceived by eye (underexposed) and detail in the darkest areas will be lost.
The lost shadow detail can't be recovered by what can be done is to lighten all the pixel values above 0.0.0 to the value they'd be in a full range image. That's what happens when a middle slider levels adjustment is use. It will "globally" lighten the mid- and 3/4 tones to the point they look similar to what was seen by eye.
The shadow detail lost at capture? In most photos the 1-3 stops of scene range seen by eye but not captured by camera don't contain critical detail and the viewer won't dwell on them unless doing a technical critique of the tonal range.
Presented with a predominantly dark background photo where will the viewer gravitate to and dwell? The lightest most strongly contrasting area. When composing photos I ask myself what I want to be the "punchline" focal point then ask how I can make it contrast and instinctively capture the viewer's attention. The fact the capture will lose shadow detail is factored into the composition decisions to create the dark > light eye path dynamic. So I expose for the highlights, normalize the mid-tones, and don't sweat a loss if detail in a scenic.
For portraits if the ambient contrast exceeds sensor range I put the sun at their back and use two flashes in a key-fill configuration just as I would with key:fill + hairlight in the studio. Set exposure to put detail in the sunny highlights on the white shirt, add chin level fill flash to put detail in the black suit, then raise intensity the key light on the front (at a flattering angle matching the skylight) until the front side of the white shirt is rendered about 1/2 stop darker than the sunny highlights (which retain detail). Sunny white on Zone 9 (250), Shaded white on Zone 8 (235-240), Black Suit on Zone 1 (30-40). The key:fill flash allows me to match foreground subject to sensor range exactly, so the result is "seen by eye" normal over the entire tonal range and there's no need for a middle slider Levels mid-tone correction in post processing as in ambient only shots.
The Catch-22 is the ambient background in the flash assisted portrait exposed to have the sunny highlights below clipping will have darker than normal mid-tones and loss of shadow detail. The solution to that problem is avoiding shooting in front of backgrounds where than other than normally expected rendering will be noticed, or using the darker rendered background intentionally to draw attention to the brighter normal contrast face.
Photos captured in overcast lighting typically have ranges shorter than the sensor range. When exposed for highlights the mid-tones and shadows in the photo look lighter and "flatter" than seen by eye. In that situation normalizing the captured image in Levels would first require moving the shadow slider right until the shadow are black and normal looking.
In Levels when you move the shadow or highlight sliders you'll see the middle slider also moves. That's PS trying to automatically normalize the mid-tones as the overall range is shortened to increase / decrease white:black endpoint contrast. I'll then move the middle slider manually back and forth for before/after comparison to see what best matches my impression in person or the mood I want it to project.
As I did back when I did Zone System B/W, once I get the normalized full range of detail in the image (which I actually do at the RAW stage in ACR) I open in PS, add screen and multiply layers and locally adjust areas of the image lighter / darker similar to dodging/burning in a print to get the desired creative results. On darker background images I'll usually darken the edges of the image, sometimes to the point shadow detail recorded by the capture is rendered nearly solid black giving the viewer a subliminal clue the stuff with the detail in the middle is what they should go look at