Newbie Feedback

[Jack Chidley]

I wanted to confirm what the power loss would be for a mechanical heat recovery fan. I had tried to do this on a calculator but this quickly became annoying: the need to check units, do unit conversions (all scalar!), revise calculations etc.

So I thought that I'd try to use rust to do the same thing. I quickly found the uom library and entered the following program:

#![allow(mixed_script_confusables)]
#![allow(unused_variables)]

use uom::fmt::DisplayStyle::Description;
use uom::si::f32::*;
use uom::si::mass_density::kilogram_per_cubic_meter;
use uom::si::power::watt;
use uom::si::specific_heat_capacity::kilojoule_per_kilogram_kelvin;
use uom::si::temperature_interval;
use uom::si::volume_rate::liter_per_second;

/// Heat loss by ventilation
/// https://www.engineeringtoolbox.com/heat-loss-buildings-d_113.html
/// Hv = cp ρ qv (ti - to )
/// where:
/// Hv = ventilation heat loss (W)
/// cp = specific heat air (J/kg K)
/// ρ = density of air (kg/m3 )
/// qv = air volume flow (m3 /s)
/// ti = inside air temperature ( oC)
/// to = outside air temperature ( oC)
/// The heat loss due to ventilation with heat recovery can be expressed as:
/// Hv = (1 - β/100) cp ρ qv (ti - to )
/// where
/// β = heat recovery efficiency (%)
///
fn main() {
    let cp = SpecificHeatCapacity::new::<kilojoule_per_kilogram_kelvin>(0.7171);
    let ρ = MassDensity::new::<kilogram_per_cubic_meter>(1.276);
    let ti = TemperatureInterval::new::<temperature_interval::degree_celsius>(20.0);
    let to = TemperatureInterval::new::<temperature_interval::degree_celsius>(10.0);
    let qv = VolumeRate::new::<liter_per_second>(15.0);
    let β = 50.0; //efficiency as percentage
    let hv = (1.0 - β / 100.0) * cp * ρ * qv * (ti - to);

    println!("{hv:?}"); // inital ouput

    // added later
    println!("{:.0}", hv.into_format_args(watt, Description));
    doesnt_work(hv);
}

fn doesnt_work(hv: Power) {
    use uom::fmt::DisplayStyle::Abbreviation;
    use uom::si::time::second;
    let s = Time::format_args(second, Abbreviation);
    // difficult to understand error
    // println!("{}", s.with(hv));

    // easy to understand error
    // println!("{}", hv.into_format_args(second, Description));
}

The initial implementation of main ended with line 36: println!("{hv:?}"); with this output 68.62647 m^2 kg^1 s^-3

Because I'm a science/engineering graduate, I recognised the output as base SI units (I knew that 68 was about right) and so looked up the definition of "watt" to confirm that it was correct.

The fact that I was able write this quickly in rust, including using the original Greek letters, and get the output in properly typed SI units is amazing. This illustrates to me why it's worth investing in rust and how great the uom library is.

There were a number of things that tripped me up and took me a long time deal with:

• "liter" is used in the uom library for the SI quantity "litre"
• ThermodynamicTemperature vs TemperatureInterval
• moving from debug println output of "m^2 kg^1 s^-3" base units to "watts" derived unit in "proper" println output
• Rust-Analyzer integration: for example it annotates SpecificHeatCapacity as Quantitly<dyn Dimension<L = ... etc
• Type annotations: the need to use Power as the type annotation (from uom::si::f32 and not somewhere else). It's difficult to write functions without this.
• error messages: try uncommenting println!("{}", s.with(hv)); in doesnt_work function and then try uncommenting println!("{}", hv.into_format_args(second, Description)); instead. See below for the compiler output of each, below.

    Checking heating v0.1.0 (C:\Users\...\heating)
error[E0308]: mismatched types
  --> src\main.rs:46:27
   |
46 |     println!("{}", s.with(hv));
   |                      ---- ^^ expected `PInt<UInt<..., ...>>`, found `NInt<UInt<..., ...>>`
   |                      |
   |                      arguments to this method are incorrect
   |
   = note: expected struct `Quantity<dyn Dimension<I = ..., J = ..., Kind = ..., L = ..., M = ..., N = ..., T = ..., Th = ...>, ..., ...>` (`PInt<UInt<..., ...>>`)
              found struct `Quantity<dyn Dimension<I = ..., J = ..., Kind = ..., L = ..., M = ..., N = ..., T = ..., Th = ...>, ..., ...>` (`NInt<UInt<..., ...>>`)
   = note: the full type name has been written to 'C:\Users\jackc\git\heating\target\debug\deps\heating-c42011534d221ce9.long-type-2057323855491684842.txt'
note: method defined here
  --> C:\Users\jackc\.cargo\registry\src\index.crates.io-6f17d22bba15001f\uom-0.35.0\src\si\time.rs:6:1
   |
6  | / quantity! {
    7  | |     /// Time (base unit second, s).
8  | |     quantity: Time; "time";
9  | |     /// Dimension of time, T (base unit second, s).
...  |
55 | |     }
56 | | }
   | |_^
   = note: this error originates in the macro `quantity` (in Nightly builds, run with -Z macro-backtrace for more info)

For more information about this error, try `rustc --explain E0308`.
error: could not compile `heating` (bin "heating") due to 1 previous error
}

    Checking heating v0.1.0 (C:\Users\...\heating)
error[E0277]: the trait bound `uom::si::time::second: uom::si::power::Unit` is not satisfied
  --> src\main.rs:49:40
   |
49 |     println!("{}", hv.into_format_args(second, Description));
   |                       ---------------- ^^^^^^ the trait `uom::si::power::Unit` is not implemented for `uom::si::time::second`
   |                       |
   |                       required by a bound introduced by this call
   |
   = help: the following other types implement trait `uom::si::power::Unit`:
             uom::si::power::yottawatt
             uom::si::power::zettawatt
             uom::si::power::exawatt
             uom::si::power::petawatt
             uom::si::power::terawatt
             uom::si::power::gigawatt
             uom::si::power::megawatt
             uom::si::power::kilowatt
           and 23 others
note: required by a bound in `power::<impl Quantity<(dyn Dimension<I = Z0, J = Z0, Kind = (dyn Kind + 'static), L = PInt<UInt<UInt<UTerm, B1>, B0>>, M = PInt<UInt<UTerm, B1>>, N = Z0, T = NInt<UInt<UInt<UTerm, B1>, B1>>, Th = Z0> + 'static), U, V>>::into_format_args`
  --> C:\Users\jackc\.cargo\registry\src\index.crates.io-6f17d22bba15001f\uom-0.35.0\src\si\power.rs:3:1
   |
3  | / quantity! {
4  | |     /// Power (base unit watt, m² · kg · s⁻³).
5  | |     quantity: Power; "power";
6  | |     /// Dimension of power, L²MT⁻³ (base unit watt, m² · kg · s⁻...
...  |
56 | |     }
57 | | }
   | | ^
   | | |
   | |_required by a bound in this associated function
   |   required by this bound in `power::<impl Quantity<dyn Dimension<I = Z0, J = Z0, Kind = dyn Kind, L = PInt<UInt<UInt<UTerm, B1>, B0>>, M = PInt<UInt<UTerm, B1>>, N = Z0, T = NInt<UInt<UInt<UTerm, B1>, B1>>, Th = Z0>, U, V>>::into_format_args`
   = note: this error originates in the macro `quantity` (in Nightly builds, run with -Z macro-backtrace for more info)

For more information about this error, try `rustc --explain E0277`.
error: could not compile `heating` (bin "heating") due to 1 previous error

[Mike Boutin]

Thanks for all the great positive feedback and explicit examples of issues!

A few of these already have issue:

• Spelling is a long-standing issue that just hasn't gotten attention. See #149.
• Print formatting documentation is another long-standing issue. See #123.
• Rust-analyzer or and kind of compiler error output is painful. #296 exists for error output. I'm not sure if this would help rust-analyzer.

A bit more feedback about the rest would be great:

• Do you have more details about your issues with ThermodynamicTemperature vs. TemperatureInterval?
• Do you have more feedback about needing to use uom::si::f32::Power? Is the need to reference the underlying type confusing? There is a type alias at uom::si::power::Power, but this is even less friendly to use.