From: MovGP0 <movgp0@gm...>  20060728 22:34:23

Hi, I've currently a lot of discussion about how to use °C (Degree Celsius) and the like for the representation of Temperaturedifference. As you may know has every Unit describing Temperature except Kelvin a nonabsolute ZeroPoint (notice the "Degree" Keyword). That means that every of these units don't represent a factor, but a function. For example the Temperature given in Degree Celsius is defined as: θ = T {θ} = {T}  273.15 where θ is the Temperature given in Degree Celsius; T is the same Temperature in Kelvin; and {x} is the numbervalue of x. In another view you can think of Celsius as a function: °C (x) = (273.15 + x) ∙ 1 K Because of this definition 5°C  2°C is not 3°C, but 3 Kelvin. Therefore we can't define the relative defined units * Degree Celsius * Degree Fahrenheit * Degree Rankine * Degree Delisle * Degree Newton * Degree Réaumur * Degree Rømer and some others like Bel (Decibel), just by defining a factor. Instead we need a factor and a constant for defining Temperatures, and maybe complex formulas for units like Decibel, rather than just a factor. Also I see troubles with componded units like "Ohmmeter" or "Meter per Second" vs. "inch per minute" which are currently not defineable to. Surly you could write dedicated UnitParsers for every physical Size like Acceleration, Frequency, Impulse, etc. But it might be more interresting to have a syntax to handle every unit in a generic way. see also: * http://de.wikipedia.org/wiki/Grad_Celsius#Definition (german; Definition of Degree Celsius) * http://en.wikipedia.org/wiki/Temperature_conversion_formulas * http://en.wikipedia.org/wiki/Decibel * http://wiki.ontoworld.org/index.php/Help:Custom_units * http://de.wikipedia.org/wiki/Liste_der_physikalischen_Formelzeichen (german; Overview of the most important physical Sizes) ys, MovGP0 