Floating point constants?

[Ken Elvehjem]

Hi,

I am trying to create a global constant for a floating point value.

I can find examples galore of Integer constants in the help files, but no references to using a floating point constant.

using a % sign is apparently indicating an integer... so this doesn't work:

%My_Constant = 33.7

I know I could just create a global variable and assign a value to it... Is that what I have to do, or is there something I have missed?

Ken

[Eric Pearson]

> using a % sign is apparently indicating an integer

Correct.

The easiest way to name a floating point value is to use a MACRO.

-- Eric

[Walt Thompson]

Floating Point Constants

Hi Ken;

There are two ways to solve this:
1. Declare a numeric constant (%337) and divide it by 10 when ever you use it.

2. Easier, declare a string constant ($33.7) and convert it to numeric whenever you use it. Example: X = VAL($33.7) where X is previously DIM'd as SINGLE.

[StanHelton]

MACRO My_Constant = 33.7

[Ken Elvehjem]

OK,

Thanks for the responses, guys. I was wondering if I'd had a dumb-out or something. Isn't it a little odd that floating point constants aren't "built-in" considering the number of physical constants and bugger factors involved in science and industry???

On the other hand, I haven't missed them in the last 10 years... guess I'll bone up on macros.

Thanks again for the replies!

Ken

[John Gleason]

Note below code demonstrating precision differences. It may have relevance.

#COMPILE EXE
#DIM ALL
MACRO mConst1 = .315       '<< okay if single precision is adequate
MACRO mConst2 = 315 / 1000 '<< if you need it to be EXACTLY .315
FUNCTION PBMAIN () AS LONG
       
       'thanks to Paul Purvis, I know 315 / 1000 = VAL(".315") which is the
       'exact EXT representation of .315 (or whatever your float pt. value is)
       'The following shows as defined, it is a single.
       ? mConst1 - (315 / 1000)
       ? mConst2 - (315 / 1000)
'       WAITKEY$

END FUNCTION

[Petr Schreiber jr]

Hi,

in PB I prefer to use MACROs for floating point constatns.
To maintain precision, I add the "##" for EXT.

MACRO mConst2 = 0.315##

But the 315/1000 way is very nice , I would just suggest to surround it with brackets:

MACRO mConst2 = (315 / 1000)

Because you could get into trouble - MACRO works like text substitution, which without brackets leads to "surprises" in case you do not pay much attention when writing code. See here:

#COMPILE EXE
#DIM ALL
MACRO mConst         = 315 / 1000
MACRO mConstBrackets =(315 / 1000)
FUNCTION PBMAIN () AS LONG

       ? STR$(mConst^2)
       ' = 0.000315 because it is evaluated like 315 / (1000 ^ 2) thanks to operator precedence
       ? STR$(mConstBrackets^2)
       ' = 0.099225 because it is correctly evaluated like (315/1000)^2
       
'       WAITKEY$

END FUNCTION

Petr

[John Gleason]

MACRO mConst2 = (315 / 1000)

Yes, good point Petr. I too would prefer the "MACRO mConst2 = 0.315##" but with the glitch noted in previous posts where it gets assigned as a DOUBLE rather than EXT, the (315 / 1000) looks like the ticket for MACRO full EXT precision.

[Michael Mattias]

I am working on that invoice printing application and wanted all the X-Y locations for literals, data and images in constants, in inches. So I just multiplied by 100 to make the constant, and divide by 100 in line as needed.

e.g.,

%TWO_INCHES = 200&
%ONE_HALF_INCH = 50&

Works in this application just ducky.

[Ken Elvehjem]

OK

Thanks for the responses and esp. the notes on text substitution... As ever there's more than one way to do it!

With appreciation,

Ken