With all the things you can do with editing software it almost makes the fundamentals and mechanics of taking the pictures obsolete. You can just take the pic any way you want and then fix it in the editing software later.
Yes and no. You need to do your part in capturing the image properly (or in the manner/style you want to present the final product). The camera can only do so much after that. There's a couple factors in this process.
Data - RAW files capture a ton more data than .jpg files. Think of the details in the image, and light as data. Would you rather work with little data or big data? Big data every time!
Dynamic Range - This is easily described as the spectrum between bright and dark. Your eye sees something like 20 stops of dynamic range. This is what allows you to see that broad difference between something bright as well as pick out the details in a shadow at the same time. But even with that great dynamic range, there's a adjustment period of several seconds to achieve that view with the human eye. The cameras of today only have about 11 stops of dynamic range and they're trying to capture that in a fraction of a second. So there's some limitations in the algorithm's used to actually create the image you see. That's where software comes into play.
The editing software in the dynamic range category allows you (if shooting RAW) to pull out those dark details to be viewable. On the flip side, highlights are actually a loss of data, ie: if it's super white, there's no coming back from that as there's literally no data there to work with.
Between photo stacking, merging or HDR functions, the more data you have to work with the better your final image could possibly be as it relates to what your eye saw. Conversely, the more 'artsy' you can get away from reality with all that data as well. That's totally the photographers choice as to what they want to present.
You can think of all this as simply data manipulation in the dynamic range spectrum. Don't like that beam on the roof? Clone it out. Again, data manipulation via a clone tool of some variety. ie: take the data next to the area and simply clone that data to where the beam used to be. (BTW - that beam would be a bit difficult due to the fading light from left to right as well as top to bottom)
Same goes for editing portraits and faces.... You're manipulating the data to remove pimples, wrinkles, stray hair and such. You can also elongate the necks, change the profile of cheekbones, smooth skin and a host of other features. In the digital world, this is all possible.
But it all rests with using as much data as you have available and your desire to create what YOU want to present. Some folks like "true to the eye", other like wild/artsy creations. That's the beauty in all this - it's your choice.