Toggle = not Toggle that is the question

[Arnold]

Hi Charles,

I would like to ask a question about boolean values. I tried out this little test code:

Code: OxygenBasic

1. 'Toggle.o2bas
2.  
3. include "$/inc/minwin.inc"
4. include "$/inc/console.inc"
5.  
6. print "True = " true & cr
7. print "False = " false & cr & cr
8.  
9. bool toggle = not false  'true
10.  
11. while 1
12.   toggle=not toggle
13.   print "toggle = " toggle & cr
14.   if not toggle then print "this is: not toggle  (" toggle ")" & cr
15.   print "Enter ..." & cr & cr
16.   waitkey()
17. wend
18.  

If I run this code the case of 'not toggle' will be displayed correctly. If I apply at the beginning: bool toggle = true then this will not work. (not toggle -> -2).

I suppose there are reasons why you use -1/0 for true/false in boolean types? But why is 1/-2 used for true/false values if I apply the true value of minwin.inc?

Roland

.

[Patrice Terrier]

This is a classical behavior.

Better to define is as long

long toggle = 0

toggle = not toggle // then toggle = -1

if toggle is define as toogle = 1, then you will always get a bad result with the NOT operator.

[Arnold]

Hi Patrice,

bool toggle
bool toggle = false
bool toggle = -1

these combinations will also work with OxygenBasic besides using long. So there would be enough workarounds. I only wondered if there is a reason for 1/-2 values when using 1 as a value for True.

Roland

[Patrice Terrier]

This is because NOT 1 = -2

[Arnold]

Thank you for the help. I did not expect this result, but my windows calculator gives the same answer. Although I did not yet find the explanation in Internet for this rule I will keep it in mind for the future.

Roland

[Charles Pegge]

Hi Roland & Patrice,

I would add that Oxygen does not use real boolean types. It uses 32 bit integers, in common with C/Windows API.

All Oxygen logic is bitwise, so when performing Boolean logic, use % true -1

some negative binaries and their end bit patterns:

-1  ffff  1111
-2  fffe  1110
-3  fffd  1101
-4  fffc  1100

[Mike Lobanovsky]

In order to follow Boolean logic rules, True doesn't have to be defined at all. With False = 0 as in all the languages I'm aware of, the assumption that True is anything that's <> False | 0 is sufficient to express every possible logical combination. The numeric representation of True as -1 for retro-BASICs and other languages with a long historical record is purely traditional and is strictly engine implementation specific; any representation would do as long as it is consistent and taken account of by the engine.

Note that C defines False as long 0, and True is in fact anything that's not False, with the internal numeric representation of True being 1. C++, C#, Visual Basic, and VB.NET follow the same pattern. FBSL follows it too.

Also note that, despite a particular numeric representation of True and False in a given language, many languages would not allow you direct arithmetics (hexa, decimal, octal, or binary) with Booleans or comparison of Boolean values for equality/inequality with numeric values where the sign of their numeric representation might have otherwise affected the result.

For the sake of compatibility with major modern BASICs and C/C++/C# and their derivatives, I'd recommend sticking with the C interpretation of Booleans when designing your own Boolean based branching in your apps.

[JRS]

Mike is correct.

Code: [Select]

TRUE = NOT FALSE


[Mike Lobanovsky]

Yes, John, and that's the only thing you can safely rely on. Everything else comes from the devil (or the dev).

[Charles Pegge]

Laws of Form by G Spencer-Brown, provides A very interesting view on the roots of logic.

https://en.wikipedia.org/wiki/Laws_of_Form

In this perspective, conflating operators and operands, True is the null no-action state, and False is the distinctive/transformative state.

[JRS]

The belief that anything other than zero is TRUE seems correct in my mind. In Script BASIC, FALSE = 0 and TRUE = -1. TRUE/FALSE are SB keywords with predefined values.

Code: Script BASIC

1. T = 10
2. IF T = NOT FALSE THEN
3.   PRINT T,"\n"
4. END IF
5.  

jrs@laptop:~/sb/sb22/test$ scriba notfalse.sb
jrs@laptop:~/sb/sb22/test$


OT

As to the discussion on BP.org about language popularity, PHP and JavaScript are the most widely used languages in todays computing. The desktop is no longer king.

[Arnold]

Thank you again for all the info. I now understand that:

not   0=-1, not  1=-2, not  5=-6, not 17=-18 etc.
not -18=17, not -6=5,  not -2=1,  not -1=0

This looks simple, but it was not for me. After some experimenting I found that False and True defined in minwin.inc should work, sometimes with a small workaround. Therefore for some special cases I will use:

% bTrue=-1 'bitwise True

I think this should be sufficient to catch most of the possible False / True cases.

Roland

[Peter]

As to the discussion on BP.org about language popularity, PHP and JavaScript are the most widely used languages in todays computing. The desktop is no longer king.

Now I am slave. 

[JRS]

Now I am slave.

Could be worse. You could be someone's bitch. 

Computed GOTO

In the day ...

Code: [Select]

ON x GOTO 100, 200, 300, 400, ...


[Peter]

Computed GOTO

will do when I'm dead.