Re: Loops (was Re: do { quit; } else { })

On 16.04.2025 16:31, bart wrote:

On 16/04/2025 13:08, Michael S wrote:

[...]

 
When languages used to be 1-based, it was easy. Then 0-based came along,
and typically iteration changed from 1..N inclusive, to 0..N-1 inclusive.

You should be aware that [simple] linear iterations are conceptually
coupled to arrays. The choice of 0-based arrays and thus iterations is
"useful" for compiler creators of such languages. They allow without
overhead and skill for optimization an efficient array access.

[...]

Oh, now you could interpret a written above as statement of superiority
of C syntax. So, no, it is not. Those are *minor* points.

 
Here's some C code to print the elements of an array:
 
    static char* table[] = {"one", "two", "three", "four", "five"};
 
    for (int i=0; i<sizeof(table)/sizeof(table[0]); ++i)
        printf("%d %s\n", i, table[i]);

(This all exposes quite clearly the low-level characteristics of "C".)

And here is how I'd write it using the 'minor' advantages of my syntax:
 
    static []ichar table = ("one", "two", "three", "four", "five")
 
    for i, s in table do
        println i, s
    end
 
To me, the differences in that for-loop are like chalk and cheese.

Sure, many language support newer syntaxes for such things. Even older
languages do such abstractions; e.g. Awk[*]

    split ("one two three four five", table)
    for (i in table)
        print i, table[i]

(One of many things I like in this very small but powerful language.)

 
(Note that the latter is 1-based, but it is not apparent in the code;
the outputs are numbered 1 to 5. The C is zero-based, and that has to be
specified. The outputs are numbered 0 to 4.
 
Moreover, I could have specified a different array base, and the loop
remains unchanged.)

Janis

[*] Read it as "GNU Awk", since in that variant (as opposed to standard
Awk) you can control the 'for' iteration order.