[sgtnoodle]

The project looks interesting. I am always a fan of embedded virtual machines. I have some questions after skimming the documentation.

Why do built-in functions use a keyword/operator syntax for arguments, while user defined functions use parentheses? That seems confusing (and was deemed a mistake in python worthy of a decade of pain to remedy.) It would certainly make sense to special case their implementation in byte code for efficiency, but since there's a compilation step, why does the syntax need to change?

Why did you choose to go with Arduino function names? Do you like the camelCasing?

Why is there both a "next" keyword and a "continue" keyword? Don't they basically do the same thing? If the purpose of "next" specifically is to mark the end of scope of the loop, why not use the word "done" or another word that isn't synonymous with "continue"?

If space and efficiency are a priority, and there's a compiler from a higher level language, why did you go with an ASCII-code rather than a true byte-code? That gives you 95 visible characters per byte, which seems to potentially waste 18% or so of program space.

[gioscarab]

Thinking about it more in detail, functions require the parentheses to show where the call ends. I obviously could count the number of expected parameters as function is defined, but that leaves room for error, and a creepy syntax,

ie: sum 2, multiply 2, 2, 2

for which function is the last parameter for?

You are right about the function names, I am not forced to use the camel form which I do not love, it could be serial_write or even serial write

[avhon1]

This is Polish Notation, and it does not have this problem as long as you don't allow variadic functions. If functions always consume the same number of arguments, it's no problem at all.

This is basically how concatenative stack-based languages do it, although by convention they put the arguments before the function call. (Reverse Polish Notation) In Forth, your example would be written one of the following two ways, depending on whether multiply() or sum() is consuming the extra 2:

2 2 2 * 2 + +

> 8

2 2 2 * * 2 +

> 10

Also valid would be:

2 2 2 * 2 +

> 2 6

where the left-most 2 is not consumed, but rather is left in place.

[gioscarab]

Ciao, yes I agree with you, I will work to remove parentheses from functions. What you would suggest instead of the arduino function names? You are right also about next and continue, thank you. The fact that the machine language is an ASCII string looks like really handy for many different reasons.

[sgtnoodle]

I actually suggest keeping the parentheses, and updating the system functions to also use them. That way, you can potentially more easily support nesting of compound expressions in function calls.

Re: naming for IO functions, it's really up to you. I personally try to avoid abbreviations for readability, except for incredibly common ones.

One idea:

init_pin(index, mode) get_pin(index) get_pin_mode(index) set_pin_mode(index, mode)

Or CamelCase, InitPin, GetPin, etc.

Mode is an enum with values for input, input_pullup, input_pulldown, output_low, output_high.

For getting timestamps:

milliseconds() microseconds()

For computing deltas in time between two timestamps correctly with rollover (Provided by the system in case the data type is unsigned. Returns 0 if t2 > t1):

time_left(t1, t2)

[gioscarab]

I really need some help on real-time operative systems, the interface for windows is more or less ready, and the code should work on win out of the box, it should be rather simple to create a console program that can:

BIPLAN compile program.txt (compiles in program.bp/bip/bipl)

BIPLAN run program.bp/bip/bipl

would be helpful and fun to have it running on windows or linux.