Strange behaviour with "Double" type

[Eddy Van Esch]

Why the complier doesn't already know it's type Double, I don't know! I just declared it!!!

The compiler first converts the ascii string "1.0005" to a binary floating point number.
If you don't specify "1.0005#", the compiler makes this a single precision float (SINGLE).
After that, it stores your SINGLE 1.0005 into a variable kk, which is a DOUBLE.
With the conversion from ascii string "1.0005" to a SINGLE, a rounding error occurs, due to the limited precision of a SINGLE .

Kind regards

[Alex Chambers]

Thanks.

Why the complier doesn't already know it's type Double, I don't know! I just declared it!!!

Anyhow, this only seems to be the case for Doubles - Singles and Longs are fine without the type specifier.

[Petr Schreiber jr]

Hi,

I am afraid you need to help compiler a bit by adding "type specifier".
So just change:

kk = 1.0005

to

kk = 1.0005#

# is symbol for double precision.


Bye,
Petr

[Alex Chambers]

Hello,

I'm finding strange things with the double float type:

LOCAL kk AS DOUBLE

kk = 1.0005
MSGBOX(STR$(kk))


This gives a msgbox of "1.00049996376038" !!!

?????

It should be "1.0005"

Why is this happening and can I correct the behaviour? I'm using PB/Win 8.0


Thanks, alex