Hi Dudes,
        I had a statement
        tax_paying := tax_paying - (30000 * 0.02);

Tax_paying as you would imagine was an integer, however, I got a type
mismatch error when I compiled the program with tax_paying as an integer,
(or Longint) if I compiled it when tax_paying was a real, the answer was
not what I expected (something like 6e000+02 came up). Can you help?

In article <66e10g$ff...@tst.hk.super.net>,
Luk Hing Yuen <wh...@is4.hk.super.net> wrote:

You cannot assign floating point values into integer variables just like
that. You need to specify truncation (trunc() or rounding (round()).
Beware that the result of rounding in case of exactly xxxx.5  depends on
whether you use coprocessor {$N+} or not.

On 7 Dec 1997 11:24:00 GMT, wh...@is4.hk.super.net (Luk Hing Yuen)
wrote:

leave tax_paying as a real number, and when you write it to the screen
or file, use something like this:

WriteLn (tax_paying:10:2);

This will get rid of the scientific notation.

Hi Jacky,

The reason for this is that multiplying by a fraction (or, in general,
by a non-integer number) is the same as dividing, which yeilds,
always, a REAL result. If you are sure that you need to get an integer
result in this place, always, then you may use Round(30000*0.02), but
if the 30000 or the 0.02 are a result of something that changes during
program execution, and thus may cause a different (and not always
integer) result, a better bet would be to declare Tax_Paying as Real.

Btw, Round is better here than any other function, because it
compensates for rounding during internel machine calculations.

Good luck,

Eytan.

On 7 Dec 1997 11:24:00 GMT, wh...@is4.hk.super.net (Luk Hing Yuen)
wrote:

wh...@is4.hk.super.net (Luk Hing Yuen) was quite brave when they
trasmitted the following message:

tax_paying:=tax_paying - INT(30000 * 0.02);

OR

tax_paying:=tax_paying - ROUND(30000 * 0.02);

In article <66e10g$ff...@tst.hk.super.net>, wh...@is4.hk.super.net says...