You might think they are all different at a superficial glance. However there are linguistic universals, which are shared by large quantities of languages, and larger still are the similarities when you take away word-order and whether something is an adposition or an affix.
So it is quite easy to use the most common and regular grammar, which was most likely spoken by mitochondrial Eve near the great lakes of Africa. Post-positional SOV is the kind of grammar we are so innately attuned to, that deaf people, never having heard spoken language develop it all on their own (https://en.wikipedia.org/wiki/Al-Sayyid_Bedouin_Sign_Language ). Even people that speak an SVO language natively do better with a new SOV language vs an SVO one http://langcog.stanford.edu/papers/TFJ-cogsci2011.pdf
Anyways, I don't expect many people to learn a new language, that why the ideal grammar is mostly reserved for the core-language or interlingua which is also the bytecode. Instead there are formal variants of human languages which share vocabulary and word order of the native language -- even though they are more regular and have a more limited dictionary.
Ha, that is actually quite misleading, since the keywords are not the bulk of the language. The bulk of the language is in the libraries, and their documentation, which use a haphazard naming convention due to the lack of any standard vocabulary or grammar.
For the SPEL dictionary, I used 38,000 of the most common English words, and translated them into 30+ of the most common human languages. Then I filtered the ones that were overly ambigious or had too many homonyms, leaving me with and orthogonal set of just over 8,000 core words -- they can be extended through compounding.
That is an unfortunate truism mostly because of the visual-spatial nature of modern programming languages, as you so aptly demonstrated with one of the most obvious offenders (Python).
I don't consider brevity to be necessarily beneficial, particularly if it does not increase understanding. A human speaking at a leisurely rate can pronounce 3 syllables per second, so even a wordy sentence can be spoken quite quickly.
Even the sEnglish snippet had many confusing and unpronounceable parts which made it quite impossible to understand by simply reading it (like whatever Pc Hd Obst Obst2 meant to the writer is completely opaque to the casual reader).
In fact in SPEL (Speakable Programming for Every Language), such confusing and unhelpful variable naming would be difficult to accomplish, since everything is composed of valid vocabulary words.
I wont attempt to make a program of what your code did, since I can't understand it due to the confusing choice of variable names, and lack of description.
Unfortunately that is likely to lead to make it overfit for English, and underfit for the 75% of people that don't speak English, and the 95% who aren't native speakers.
Though your example is somewhat easier to understand, it still uses visual-spatial queues -- thus can't be spoken. Also it has comments which are in a different language.
The comments and error messages in SPEL are all written in SPEL, so depending on locale can be output appropriately.
I'm currently implementing a SPEL to compile to OpenCL C, since an issue people often complain about with natural language based programming langauges is that they are domain specific, high level, and-or slow. C is general and low level, wheras OpenCL harnesses the cheap TFLOPs.
here is my current hello world program (c is a "sh" sound):
hromta hnimna cwasli kratta cwasli tyeh hwacwu tyehka hsactu zrondofi fe
formated for your visual-spatial benefit:
hromta hnimna cwasli
tyeh hwacwu tyehka hsactu
in SOV English:
program _topic_case name _nominative_case begin _realis_mood
cardinal _topic_case begin _realis_mood
_text world _vocative_case _text _accusative_case say _deontic_mood
zero _number _return
in SOV English with abbreviated grammar:
program _top name _nom begin _rea
cardinal _top begin _rea
_text world _voc _text _acc say _deo
zero _num _ret
// get locale
const char hwacwu = translate(locale,"hwacwu"); // get translation
/* hidden is the translations generated in the background, .po files or otherwise */
in SVO English with punctuation:
about program, name, begin.
about cardinal, begin.
say, text hey world text!
number zero return.
The result of the program in an English locale would be "hey world".
In Middle English it might be "O world",
In a Pyash locale it would be "hwacwu" (hwashwu).