Widening

[donal horgan]

Can someone please explain to me why going from a long(64bits) to a float(32 bits) is considered widening?

[marc weber]

In short (no pun intended), the approximate range of a long is plus or minus 9.2 X 10^18, but the approximate range of a float is plus or minus 3.4 x 10^38.

Note the magnitudes here: Ten to the 18th for a long, compared to ten to the 38th for a float. In other words, the 32-bit float has a vastly greater range than the 64-bit long.

The reason is that floating-point primitives (Java's float and double) sacrifice precision for range.

See this thread for detailed explanations:
http://www.coderanch.com/t/247234/java-programmer-SCJP/certification/method-parameter

And see this site for more details on IEEE 754 standards:
http://www.public.iastate.edu/~sarita/ieee754/homepage.html
[ March 18, 2005: Message edited by: marc weber ]

[donal horgan]

Cheers for the link.
Now i get it.

[Krishnakumar]

Hi,

Thanks a lot for the posting...But why is the code below throwing a compiler error..? It says integer is too long. But a suffix of l or L solves the problem.

float f = (long)9223372036854775807;

I got a bit confused because to convert a double leteral into a float you cd either suffix it with a f or F or cast it with a (float)<double literal>

Thanks!

[Mike Gershman]

From the Java Language Spcifications:

The largest decimal literal of type int is 2147483648 (231). All decimal literals from 0 to 2147483647 may appear anywhere an int literal may appear, but the literal 2147483648 may appear only as the operand of the unary negation operator -.

The expression "(long)9223372036854775807" says to create an int with the value 9223372036854775807 and then cast it to type long. But 9223372036854775807 is greater than 2147483647, so it is not a legal int literal.

This will work because there is no attempt to create an int first:
float f = 9223372036854775807L;
[ March 21, 2005: Message edited by: Mike Gershman ]

[Krishnakumar]

Thanks a lot Mike!!