 |
Author |
Topic: little suggestion on operations reference (Read 1719 times) |
|
lunetta
 
 
|
little suggestion on operations reference
« on: Apr 28th, 2004, 11:18pm » |
|
Hi
It would be useful to computational-challenged people (like me) the addition of reference on datatype conversion/preservation during operations, just like this:
http://pike.ida.liu.se/docs/tutorial/expressions/complex_types.xml
example:
int a = 5;
int b = 2;
float c = 5/2;
the actual result is 2.0; but makes people wonder "why not 2.5?"
answer: int / int = int...
|
|
|
|
|
kevinP
|
Re: little suggestion on operations reference
« Reply #1 on: Jun 1st, 2004, 11:43am » |
|
Hi lunetta,
on Apr 28th, 2004, 11:18pm, lunetta wrote:Hi
It would be useful to computational-challenged people (like me) the addition of reference on datatype conversion/preservation during operations, just like this:
http://pike.ida.liu.se/docs/tutorial/expressions/complex_types.xml
example:
int a = 5;
int b = 2;
float c = 5/2;
the actual result is 2.0; but makes people wonder "why not 2.5"
answer: int / int = int...
|
|
There is this:
http://processing.org/learning/examples/datatype_conversion.html
But this still trips me up sometimes; for example, I am not sure about the difference between 'c' and 'd' below...
Code:
int a = 5;
int b = 2;
float c = a/b; // ???
float d = float(a/b); // processing data type conversion
float e = (float) a/b; // cast
println(c); // "2.0"
println(d); // "2.0"
println(e); // "2.5"
|
|
Working right to left I would have thought that Java's "cast" would function the same as Processing's data type converter "float()" (or vice versa). Perhaps there needs to be a "see also" reference in the data type sections of the reference page.
-K
|
| « Last Edit: Jun 1st, 2004, 11:48am by kevinP » |
|
Kevin Pfeiffer
|
|
|
TomC
|
Re: little suggestion on operations reference
« Reply #2 on: Jun 1st, 2004, 1:01pm » |
|
Short answer:
Where in Java you would do:
Code:
float e = (float) a/b; // cast
|
|
In Processing this can be written as a function-style cast where it is much clearer what is happening:
Code:
float e = float(a)/b; // cast
|
|
Both of these examples can be read as "convert a to a float and then divide by b".
|
|
|
|
|
TomC
|
Re: little suggestion on operations reference
« Reply #3 on: Jun 1st, 2004, 1:04pm » |
|
Long answer:
The conventional (C-style) cast operator in Java has high precedence. That means it gets evaluated before arithmetical operators (+,-,/,*), which is why the cast happens before the divide.
Only array indices, method/function calls and member access have higher precedence than the cast operator. That means that you can cast a member of an array, the return value of a function, or a member variable of an object, without using brackets:
Code:
// Java:
// We need to know that the [] has higher
// precedence than the cast, so we're not casting
// the array, we're casting a member of the array.
float[] a = { 1.1, 2.2, 3.3 };
int b1 = (int)a[0]; // set b1 to 1,
// We need to know that function calls have higher
// precedence than the cast operator, so we're not
// casting the function, we're casting its return
// value
float myFunction() {
return 2.2;
}
int b2 = (int)myFunction(); // set b2 to 2
// We need to know that member access has higher
// precedence than casting, so we're casting the
// member variable, not the object itself
class MyClass {
float a = 3.3;
}
MyClass myClass = new MyClass();
int b3 = (int)myClass.a; // set b3 to 3
// Processing:
// We know that things inside brackets (and function
// calls) will be evaluated first, so the Processing
// cast is clearer.
int b4 = int(a[1]); // set b4 to 1
// we know that inside the brackets gets evaluated
// first
int b5 = int(myFunction()); // set b5 to 2
// We know that things inside brackets are evaluated
// first, so we're casting the member variable, not
// the object itself
int b6 = int(myClass.a); // set b6 to 3
|
|
So, to cast the result of a division, you need brackets:
Code:
// Java
float e = (float)(a/b); // cast result of int division
|
|
or a function-style cast like you used for 'float d':
Code:
// Processing
float d = float(a/b); // cast result of int division
|
|
Hope that helps.
You might find this section of the Java glossary site useful too...
http://mindprod.com/jgloss/precedence.html
And this part of the Sun Java site...
http://java.sun.com/docs/books/tutorial/java/nutsandbolts/expressions.ht ml
|
| « Last Edit: Jun 1st, 2004, 1:11pm by TomC » |
|
|
|
|
kevinP
|
Re: little suggestion on operations reference
« Reply #4 on: Jun 1st, 2004, 5:53pm » |
|
Tom,
Thanks! I see now (for the 1st time) that my biggest problem was:
Code:
// not what I wanted...
float d = float(a/b); // cast _result_ of int division
float d = float(a)/b; // what I _should_ have done
|
|
This came up before in a somewhat different form, but I hadn't made the connection that only the dividend needs to be converted. Guess I was sick on "integer math" day in school. 
Thanks again!
|
Kevin Pfeiffer
|
|
|
REAS
|
Re: little suggestion on operations reference
« Reply #5 on: Jun 12th, 2004, 10:05am » |
|
Thank you for the suggestions. I hope to be revising the reverence over the summer and will include this additional information. The ultimate goal is to have a system like PHP.net where users can append the reference.
|
|
|
|
|
|
