Hi Joanna, this pic reached me in the middle of the night, can not sleep and was fascinated in a certain way. I appreciate your skills and work since years and now you asked and I can not hesitate to start the discussion. Hope my English would be sufficient. To add: Iâm not talking DXO and what it can do to help us but purely photographyâŠ
First things first: this capture is a masterpiece in terms of perspective, technically thereâs nothing to discuss about. But to me thereâs a BUT: let me start in a philosophic way: those railways lead to anywhere and this anywhere must be found in the background â whatever it Is â but this background is here by far underrepresented, far more meaningless than the foreground in terms of sharpness and contrast. And even the perspective would need a small correction: the top left railways end at about 25% of the picture hight which is a bit problematic in terms of âGoldener Schnittâ, In total I would have preferred to go half a step back and about 30cm deeper to bring the lower left grass more to the attention and to end the top left railway 1/3 to 2/5 of picture hight. The transition in sharpness and the overall contrast should be far more homogenous, this would help the entire composition. Just my 50 Cents 
Okay, 2nd pic, a bit unlucky perspective, the middle of the doorway is the middle of the pic, I would have panned it a bit to the right and bottom to change proportion and get maybe a small step forward and I would have used a flash reduced to ÂŒ or 1/2 to push the shadows a bit because the right door seems to be interestingâŠ. You asked
Ad #1)
I think the ârule of thirdsâ and its variants came by observation of certain types of paintings. If you must have a âGoldener Schnittâ rule applied somewhere, itâs in defocus boundary, whatever that is 
Otherwise itâs about âunknownâ ahead of us, a bit of melancholy, geometry and music. The overall harmony between various aspects is more important than âGoldener Schnittâ in this case, imho.
Ad #2)
The ivy makes a story, so I like it this way. Of course, we all see different things, and you are both rightâŠ
Many folks, mostly in photo clubs look to composing according to ârulesâ.
But the first âruleâ of composition is that there are no rules. There are compositions that some folks find more pleasing for certain subjects.
I rarely compose for âthirdsâ but I use leading lines a lot. Sometimes certain layouts simply âappearâ, like this Fibonacci curveâŠ
A lot of the best composition tends to be by instinct and doesnât need measuring.
Finally, have you heard the saying, rules are there to be broken?
I knew a reporter who had the rule: âAlways have a pissing dog on your photoâ. Sometimes he used this rule literally
Yet another rule:
Create the rules after creating your photos.
Folks, please not too much attention to this rule thing, it was just an example, the other aspects are at least as important to me. And donât forget, this is not a beauty contest here, weâre just chatting. But I love to exchanging ideas with the best and Joanna asked so to sayâŠ
My background and part of this motivation is: Iâm coming from a time when it was essential to do things right at the first place, where e.g. perspective and (partial)exposure-corrections are to be made in the darkroom, when a roll of Kodakchrome25 cost 16DM and I got 4DM pocketmoney a week. And I had no lightmeter before I was 16. So I was dependent on and used to learn before I took any picture. And I loved it, later at 17 with the Leicas, Retinas and Contaxâ of my dadâŠ
Today we have it far easier with DXO and I love it since 2009. But that doesnât change my photographic genes. And by the way: I never was a member of a photoclub
Sure, and thanks for provoking the chat . Iâm not even BA, just a MINS (man in the street), exchanging âintelligent soundsâ for fun. Absolutely no offense meant, to be explicit.
Maybe AI has already discovered some of the rules our brains abide to? I often do cropping instinctively, probably having some meaning, colors, lighting in mind, the target being some kind of âbalanceâ. So it could be quite easy for AI to discover these ârulesâ, perhaps leaving the problems with emotions and context aside The problem is, thereâs often no single âbestâ outcome and too many people to judge.