i've read the first few paragraphs and i can see why HOI may have objected.
is creativity logical? i'd say, "not always."
however, when it comes time to write down the results, the proofs should be based on some form of logic though not neccessarily a strict, machine-level, programming language.
my suspicion is that even creativity and lateral thinking is still logical just super high-level
in terms of higher-level languages, we have the following structure:
the kernel of understanding
the possibly totally logical language of creativity
the language of advanced proofs (where the vast majority of the words are difficult to concisely define in a few words)
the language of proofs (more details are spilled out here and one need not be an expert in the field to accept the proof)
the formal proof language (machine language)
binary code.
the problem is that when a higher level language is projected onto a lower level one, the lower level one is almost always, if not always, longer and more complicated. the debatable question is at what level should we communicate results? we should never use either extreme, that much is clear. in other words, i can't publish in a journal: my understanding is this, period, and here are my results. and we can't publish binary code and expect a human to understand it though in order to ever get computers to prove things and/or do proof theory, then a lower level language ought to be used. what i don't know for sure is exactly what language one ought to use for proofs.
think about it: there is no such thing as a proof, ever. mathematical proofs, at least, are just words. how many verbs can you insert into the sentence so that it makes sense: "words _____?" can words love? can words kill? can words do anything? can words prove?
"proofs" do one of the following things: pass the set criteria for rigor and correctness, not, or something in between. someone writes down a proof but a human has to judge it for correctness. the words in themselves don't prove; it is the reader and writer who prove.
"proofs" are an attempt to articulate that kernel of understanding and stimulate the reader's kernel so that they will also understand. once the reader has matched the kernel of understanding enough, by a preset criteria, then the reader will judge the "proof" to be "valid."
the bottom line is that if the language is too high for humans, that kernel of understanding just won't get transferred. it's not a matter of "should it" or "could it": it typically doesn't unless the reader already has that kernel.
what seems to be the most efficient technique for transferring that kernel of understanding is for the author to write in a medium level language. but when that author talks to his friend who's also an expert, listen to their conversation. it's at a much higher level language and they can prove things to each other with much more efficiency.
i use the word "kernel" on purpose so that it resembles actual computers. if we are self-aware mathematical structures, then we might literally have kernels. i wonder if our kernels are "written" in a high level language or a low level language or something in between...