Are System.Double operators overloaded for epsilon?

[David Veeneman]

Are System.Double operators overloaded to perform epsilon comparisons, or do
these comparisons have to be performed by the programmer. In other words, if
I have two doubles, A and B, and I test "A > B", does C# do a strict
mathematical comparison, or does it do a "A - B > Double.Epsilon"
comparison? Thanks

[Nicholas Paldino [.NET/C# MVP]]

David,

I don't believe that epsilon is used in comparisons. If you need to
assure a degree of precision, then use the Decimal class, which will
guarantee precision (and therefore, remove the need for epsilons).

Hope this helps.

--
- Nicholas Paldino [.NET/C# MVP]
- mv*@spam.guard.caspershouse.com

"David Veeneman" <da****@nospam.com> wrote in message
news:ub**************@TK2MSFTNGP14.phx.gbl...

Are System.Double operators overloaded to perform epsilon comparisons, or
do these comparisons have to be performed by the programmer. In other
words, if I have two doubles, A and B, and I test "A > B", does C# do a
strict mathematical comparison, or does it do a "A - B > Double.Epsilon"
comparison? Thanks

[Octavio Hernandez]

David,

I believe the equality comparison is based on the binary pattern of the
operands, except for special cases (NaN).
It is programmer's responsibility to check against an Epsilon instead of
using '==' on floating point values.

Regards - Octavio

"David Veeneman" <da****@nospam.com> escribió en el mensaje
news:ub**************@TK2MSFTNGP14.phx.gbl...

Are System.Double operators overloaded to perform epsilon comparisons, or
do these comparisons have to be performed by the programmer. In other
words, if I have two doubles, A and B, and I test "A > B", does C# do a
strict mathematical comparison, or does it do a "A - B > Double.Epsilon"
comparison? Thanks

[Jon Skeet [C# MVP]]

Nicholas Paldino [.NET/C# MVP] <mv*@spam.guard.caspershouse.com> wrote:

I don't believe that epsilon is used in comparisons. If you need to
assure a degree of precision, then use the Decimal class, which will
guarantee precision (and therefore, remove the need for epsilons).

Decimal doesn't guarantee precision - it certainly doesn't prevent
rounding errors etc. It just changes the set of numbers which can be
exactly represented.

--
Jon Skeet - <sk***@pobox.com>
http://www.pobox.com/~skeet
If replying to the group, please do not mail me too

[Christoph Nahr]

They are not. IMO it's a good thing that they aren't since that would
be counter-intuitive, but I certainly would applaud an additional set
of standard methods to perform epsilon comparisons on floats &
doubles. I asked David Notario but I don't think he'll do it. :)

On Tue, 16 Aug 2005 14:57:46 -0500, "David Veeneman"
<da****@nospam.com> wrote:

Are System.Double operators overloaded to perform epsilon comparisons, or do
these comparisons have to be performed by the programmer. In other words, if
I have two doubles, A and B, and I test "A > B", does C# do a strict
mathematical comparison, or does it do a "A - B > Double.Epsilon"
comparison? Thanks

--
http://www.kynosarges.de