1. Sequence initialisers for "for" loops
- Posted by Lmailles at AOL.COM
Jun 30, 1998
for a=b to c by d do
why can't for loops take sequences for a, b, c and d?
for a={0,0,0} to {3,6,9} by {1,2,3} do
standard sequence matching rules apply. b,c and d must have the same length
--------------------------
Daniel Johnson
Les Mailles Technologies
"If you don't stand for God, you won't stand at all" -Isaiah
2. Re: Sequence initialisers for "for" loops
Lmailles at AOL.COM wrote:
> for a=b to c by d do
> why can't for loops take sequences for a, b, c and d?
> for a={0,0,0} to {3,6,9} by {1,2,3} do
> standard sequence matching rules apply. b,c and d
> must have the same length
IMHO: a worthy, logical, intuitive, clean, elegant and
needed addition to Euphoria's interpreter. The core
of the language is sequences and their manipulation.
Using sequences in this manner, in loops, is natural.
My unregistered (as soon as I sell a program, I'll
register :) ) vote is a definitive and emphatic yes.
take care--Hawke'
3. Re: Sequence initialisers for "for" loops
- Posted by Ralf Nieuwenhuijsen <nieuwen at XS4ALL.NL>
Jul 01, 1998
-
Last edited Jul 02, 1998
>Lmailles at AOL.COM wrote:
>> for a=b to c by d do
>> why can't for loops take sequences for a, b, c and d?
>> for a={0,0,0} to {3,6,9} by {1,2,3} do
>> standard sequence matching rules apply. b,c and d
>> must have the same length
>
>IMHO: a worthy, logical, intuitive, clean, elegant and
>needed addition to Euphoria's interpreter. The core
>of the language is sequences and their manipulation.
>Using sequences in this manner, in loops, is natural.
>My unregistered (as soon as I sell a program, I'll
>register :) ) vote is a definitive and emphatic yes.
>
>take care--Hawke'
Yes, a brilliant and powerfull adition, wich is actually even more
consistent that the present situation. I do find it ackwards at times to see
when and when not things can work with objects as well. Consider this:
type pos_object (object x)
if x < 0 or not integer (x) then
return FALSE
else
return TRUE
end type
end type
It doesn't work as it is suppose to. Off course if it did, this would no
longer work:
function peek (object x)
if integer (x) then
.. ok.. no code could be here.. but its about the logic
else
.. same here
end if
end function
I know it is too late to change, but sometimes I wonder if we would need a
symbol or something to show that we want it to recursive, and as many times:
<<a>> .. means recurse with sequence containing atoms (1D seq-thus)
integer ( <a> ) .. recurse with atoms
integer ( a ) would work as it normally does..
However..
<a> + <b> would be the new form of a + b
So, you see, it is more clear, but too late unfortunately.
No need to have different versions of Euphoria, but if there ever is gonna
be a non-compatible version of Euphoria, keep this in mind 
Maybe we can partly have this with the normal boolean logic as some one else
once brilliantly suggested:
a * b -- Normal logic, we multiply all atoms
a ** b -- New logic, we multply per 1D sequence
a *** b -- New logic, per 2D sequence..
a // b -- A 1D-seq divide
And off course the few builtin functions that don't recurse can have a new
recursive function:
integer_recurse ( a )
However, these could just be in an include file or something.
Yes, its this impulse for consistency I keep having, but this actually adds
a powerfull new trick, and won't break code, unlike the more preffered
syntax with the << & >>'s.
Ralf Nieuwenhuijsen
nieuwen at xs4all.nl
