From: "Geo" 

double a, b;

  a=15/2;
  b=15.0/2;

  cout << a << " " << b << "\n";

output is 7 7.5

Geo. (seems like a significant difference to me )

"Paul Ranson"  wrote in message
news:417a5a3f{at}w3.nls.net...
> I don't think you ever have to write '2.0' unless it's important to know
> that this is a floating point type for compile time arithmetic or function
> overloading
>
> For instance,
>
> double d1 = 22 / 7 ;
> double d2 = 22.0/7 ;
> cout << "d1 = " << d1 << ", d2 =
" << d2 << "\n" ;
>
> produces,
>
> d1 = 3, d2 = 3.14286
>
> whereas,
>
> d1 = 22 ;
>
> d2 = 7 ;
>
> std::cout << "22/7 = " << d1/d2 << "\n" ;
>
> produces,
>
> 22/7 = 3.14286
>
> All languages that do integer and floating point arithmetic have to have
> conventions for this type of thing.
>
> Paul
>
> "Geo"  wrote in message
news:4179884a{at}w3.nls.net...
> > "John Beckett"
 wrote in message
> > news:4178e2cc.15036661{at}216.144.1.254...
> >
> >> C++ uses the C philosophy that if the hardware can do it, then the
> >> language should be able to call it. So, since most hardware can handle
> >> single-precision and double-precision floating point numbers, the
> >> language
> >> should handle both.
> >
> > Ok, that's a logical philosophy.
> >
> >> If you ever wanted to work with arrays of (say) 100 million floating
> >> point
> >> numbers, you might be glad you can specify float (4 bytes in VC) rather
> >> than double (8 bytes).
> >
> > Yeah I can see that too, but there is other stuff that bothers me about
> > the
> > language. For example there are times when you have to say something is
> > =2.0
> > instead of =2 but when it cout's a double number with no output
> > specification the default is to just output 2 not 2.0. It's kinda like
the
> > machine doesn't have to play by the same rules the programmer has to
play
> > by..
> >
> > Geo.
> >
> >
>
>

--- BBBS/NT v4.01 Flag-5