I bet a lot of my programmer friends are going, "why didn't you just use JSON and serialize settings that way?" Y'all, I did consider good ol JSON, because yes, it does have a faster compute time, but before the phoneme editor and NVDA driver settings, JSON would have been a lot less human-readable. YAML does suck for indents, I agree, but so does Python, if you are thinking on those lines, and good code indenting is honestly just good code hygiene if you think about it, why not keep that alive.
I read them on a Braille display a lot, I recognize not all people do, but poorly indented code in Braille VS well-indented is the difference from knowing quickly where functions begin VS trudging code as though it were a thick foggy swamp. And I don't like that feelin'.
@Tamasg IMHO enforced indention is bad. 1st, presentation and content must be decoupled. the web has truly proven the benefits therein. 2nd, this is ableist. Many folks need things to look a certain way for their disability, be it visual, processing, attention, what have you. Enforcing indention is harmful to those folks. 3rd, this removes choice. JSON can be indented, auto or otherwise, but YAML must be, which means no choice. Tech should work for us, not the other way around.
@sinabahram but I will say: I do see where it could be seen as ablist, in the sense that someone might not have pitch perception, nor Braille, then what? But that's where I think we just need more alternatives for indicating that, haptics, ETC. Would be cool to see an IDE where it did use certain haptic bumps because that I think could even benefit everyone equally for quickly feeling a count of indent levels. I personally see it as a part of semantics, even for machines, because they can parse the number of indents and understand blocks of code better. Without it, chunking would become a lot harder for machines, too. So it benefits both worlds, it's just a matter of equally making it available to everyone.
@Tamasg I feel you're stating indention is needed. I fundamentally disagree. That is simply one way of conveying scope. Another is braces, another is fixed columns from back in the fortran days, and much more. Instead of baking inaccessibility in, let's simply not do that. Also the way it is done is itself inaccessible by enforcing spaces per Python's pep-8 instead of tabs. More importantly, nothing prevents indention being automatically added with all these other options.
@Tamasg also, I don't mean to be all computer sciency, but it is 100% not semantics. It is syntax being conflated with semantics, which is my objection. Syntax should be syntax and semantics should be semantics. It's not about opinions, but rather about data. No peer-reviewed programming language literature to the best of my knowledge supports indention making a statistical difference for programmers. So, in the abscence of evidence, I always choose choice.