In all versions of VB, why does True evaluate to negative 1?

Hello,

Ari Bixhorn, Project Manager of Visual Basic at Microsoft, mentions
coercing the value of True to be -1 as a matter of convention. It's
just the way Microsoft decided True would evaluate in VB 1.0 or
whatever the first version was. Their convention is causing a bit of
headaches in their .NET initiative, because other languages define
True to be 1 and False to be anything else other than 1, or False to
be 0 and True as anything else. That's just the way those respective
languages have always done it. There's no real rhyme or reason.
They're just numeric constants used to represent the boolean ideas of
yes or no, right or wrong, true or false. He mentions in the latest
version of Visual Basic (their .NET flavor) passing a True within the
program will send a -1, but between languages it will pass a 1 to play
nice with other languages. Ari then states at the bottom of the post:

"
Language consistency, on the other hand, implies that there is a
*right* way and a *wrong* way to declare arrays, coerce the value of
True, etc.  Neither C++ nor VB are "wrong."  They are simply
different.
"

A good programming technique is to never test for 0, or -1, or 1 for
that matter, but to use the proper boolean values TRUE or FALSE in
your programming language, that way, you can be sure no matter how
it's represented internally, your logic will flow as planned. Thank
you for your question.


Search Strategy:

"visual basic" "value +of true" on google groups:
http://groups.google.com/groups?q=%22visual+basic%22+%22value+%2Bof+true%22


Additional Links:

Confirmation of VB.NET Language Changes by Ari Bixhorn posted to
microsoft.public.dotnet.languages.vb:
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&threadm=u9SvHSrvAHA.2020%40tkmsftngp04&rnum=7&prev=/groups%3Fq%3D%2522visual%2Bbasic%2522%2B%2522value%2B%252Bof%2Btrue%2522%26hl%3Den%26lr%3D%26ie%3DUTF-8%26oe%3DUTF-8%26selm%3Du9SvHSrvAHA.2020%2540tkmsftngp04%26rnum%3D7


Thank you for the opportunity to answer your question, if you require
more information, please clarify the question, or if you find this
answer satisfactory, please feel free to rate it. Thank you!
   
skermit-ga

Thank you very much.  That is a great answer.

At the bit-level on the x86, negative numbers are encoded using a
method called "two's compliment". With this notation, negative one is
stored by setting all the bits in the number to '1'. For example, if
you have a 16-bit signed integer, it would be stored internally as 16
bits (or 2 bytes). Negative one would therefore be 16 set bits or
(1111 1111 1111 1111). The x86 architecture (and most architectures in
fact) have a NOT instruction built it which inverts the bits. Thus the
patters 0101 would become 1010 when the NOT instruction is executed on
them. When storing a boolean value, visual basic uses a byte (8 bits).
This byte is set to 256 (1111 1111) when true and 0 (0000 0000) when
false. Since all values in VB are signed, the 256 would be treated as
negative one in two's compliment. Most languages have two sets of
operators: bitwise and logical. If you store true as positive one
(0000 0001), and NOT that value, you'd get negative two (1111 1110).
Since in VB, false is 0 and all other values evaluate to true, both 1
and -2 would be true, thus there would not be a way to do this:

WHILE NOT EOF(1)
...
WEND

If EOF were false (not at the end of file), the NOT would convert it
to TRUE and the loop would execute. If EOF were true (at the end of
the file) and TRUE were positive one, the NOT would turn positive one
into negative true and the WHILE loop would consider the negative two
to be equivilent to TRUE and execute anyway, resulting in a INPUT PAST
END OF FILE error if you were reading information from the file in the
loop.

In C, TRUE is considered positive one by convention. To deal with this
fact, C uses "logical operators" to test the true/false-ness of data
and "bitwise operators" to manipulate the bits in a number.

Negative one, however is the most logical interprettation of TRUE on
platforms such as the x96 since use two's compliment notation for
storage of signed values.

-Robert

In Basic (not only in VB) logical operators (AND, OR, XOR, NOT etc.)
have an somehow ambiguous meaning.

They are in effects binary operators, not real logical operators

Binary operators perform an operation between each bit of the
operands, while logical operators perform operations between numbers
in their whole.

The problem is that, as Basic does not have a specific type for
handling logical values, we are forced to use 16-bit numbers AS if
they were logical values. This works nicely if we always work with the
pre-defined True and False values, but gets trickier when you work
with numbers.

For example, in common-sense reasoning you can say that if something
is true, it's opposite is false

This is true for logical operators but not always true with VB's
bit-wise operators.
Let's see how the NOT operation works on a number like 5. 5 in binary
is 101 ( 0000 0000 0000 0101 in 16-bit representation). Let's invert
each bit, we get 1111 1111 1111 1010. In binary two's complement
representation that makes –6.

If you try in VB: 

Debug.Print  Not 5

That's exactly what you will get. I won't enter into details of why
this happens, the important is to understand  that it happens.

The assumption " if something is true, it's opposite is false" must be
therefore be used with caution in VB programming when you use numbers
instead of the pre-defined True/False values

Try the following snippet of code:

a = 5
If a Then
  If Not a Then
   MsgBox "What? I should never get into here!"
  Else
   MsgBox "All works as we expected"
  End If
End If

A is 5, so the test "If a" succeeds, and execution continues.

But "Not a" evaluates to –6, not to 0 (the numerical equivalent of
False) so the test "If Not a" also succeeds.
 
There is an "If Not a" test within a "If a" test and both succeed!!

So, it could seem that in this case a is both true and false. The real
fact is that in this case a is neither True neither False (not the
difference in capitalisation I am using here!)

Another example of odd VB code:

a = 5
If a Then MsgBox "A IS TRUE"
If Not a Then MsgBox "A IS FALSE"

This code first says "A IS TRUE" and then "A IS FALSE".

This is just to confuse a little bit the things: all works wonderfully
if we only work with values of True and False , or their numerical
equivalents –1 and 0.

Frox-ga

In my prevoious post, there was a particularly confusing typo. The phrase

(not the difference in capitalisation I am using here!)
 
should be read:

(note the difference in capitalisation I am using here!)

Sorry for that.
frox-ga