Negative zero doesn't exist (in mathematics) so it was a mistake to giveDoesn't the IEEE spec require that -0=+0 if a compare is done?
it a representation in IEEE 754: zero should have a single representation.
Negative zero doesn't exist (in mathematics) so it was a mistake to give
it a representation in IEEE 754: zero should have a single representation.
However, when working with approximate values it may be useful
to know whether the true value is positive or negative even
when for other values a little more or less is insignifcant.
One such case is temperature where the difference between -0.1
°C can be significantly different from +0.1 °C for practical
purpoises even when the difference between, say, 5 °C and 6 °C
isn't.
On 30/05/2025 09:16, Mikko wrote:
<snip>
However, when working with approximate values it may be useful
to know whether the true value is positive or negative even
when for other values a little more or less is insignifcant.
One such case is temperature where the difference between -0.1
°C can be significantly different from +0.1 °C for practical
purpoises even when the difference between, say, 5 °C and 6 °C
isn't.
Practical porpoises will avoid -0.1°C because it makes swimming too difficult.
On 5/30/25 5:09 AM, Richard Heathfield wrote:
On 30/05/2025 09:16, Mikko wrote:
<snip>
However, when working with approximate values it may be useful
to know whether the true value is positive or negative even
when for other values a little more or less is insignifcant.
One such case is temperature where the difference between -0.1
°C can be significantly different from +0.1 °C for practical
purpoises even when the difference between, say, 5 °C and 6 °C
isn't.
Practical porpoises will avoid -0.1°C because it makes swimming too
difficult.
But porpoises swim in salt water, which will still be liquid at that temperature, but colder than they like.
Minus 0 in IEEE is mostly an artifact of the representation, being a
sign + magnatude representation.
Op 30.mei.2025 om 16:10 schreef Richard Damon:
On 5/30/25 5:09 AM, Richard Heathfield wrote:
On 30/05/2025 09:16, Mikko wrote:
<snip>
However, when working with approximate values it may be useful
to know whether the true value is positive or negative even
when for other values a little more or less is insignifcant.
One such case is temperature where the difference between -0.1
°C can be significantly different from +0.1 °C for practical
purpoises even when the difference between, say, 5 °C and 6 °C
isn't.
Practical porpoises will avoid -0.1°C because it makes swimming too difficult.
But porpoises swim in salt water, which will still be liquid at that
temperature, but colder than they like.
Minus 0 in IEEE is mostly an artifact of the representation, being a
sign + magnatude representation.
Other systems have even more possibilities to represent 0. E.g., pencil
and paper can be used to represent 0 as 0, 0.0, 0.00, 000, 00.0, 0.0E0,
-0.0, etc. It is not uncommon that a single number can be represented
in different ways in certain systems. Usually, nobody makes a problem
of it, as it is understood that always the same number is represented.
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 496 |
| Nodes: | 16 (2 / 14) |
| Uptime: | 59:03:58 |
| Calls: | 9,760 |
| Calls today: | 1 |
| Files: | 13,742 |
| Messages: | 6,185,443 |