Sciutil

A library for scientific computation.

Features

• rounding: Rounding floating-point values to match floating-point uncertainty values to one or two significant figures.
  • E.g., rounding::round_with_uncertainty(1024.0511231255, 0.015555312, "g") -> 1024.051 g ± 0.016 g.
• units: Traits and wrapper structs for treating floating-point values as physical measurements.
  • Traits and structs that embed physical units at the type level.
  • A struct that pairs a measured value with its absolute uncertainty.
• statistics: List operations for statistics.
  • statistics::derivatives: A few forms of numeric derivatives.
• display: Miscellaneous facilities for pretty-printing things.

Stability

Until it is published on https://crates.io, this library make no guarantees of stability. This is code I will write, update, and break whenever I feel the need or desire to. That being said, you are encouraged to request features or make contributions!

A part of these breaking changes includes possible name changes, if "sciutil" is ever to be registered by someone else or I otherwise decide there's a better name.

MSRV

sciutil supports only the latest stable Rust. Older version may work, but are not tested. MSRV naturally being bumped as new stable versions release is not considered a breaking change.

Looking forwards

There's a few features I'm looking to add in the future. I'm not certain if I will ever add them, but feature requests and pull requests are welcome.

• Numeric integration.
• Gate-to-gate timing (to get velocity out of photogates).
• Calculating a best-fit line with a linear regression is something I want, but I am not familiar enough with linear algebra to do that easily, as it's not something I understand. As of right now, I just use Desmos's built-in implementation.
• A simple and opinionated API for generating plots from data is also something I'd like. As of right now, I use Desmos plots, but Desmos can be limiting. Specifically, being limited to one-dimensional arrays and not having sciutil's rounding implementation are limiting for me. I'm hesitant to do it because I know it will involve a lot of work.

Documentation

sciutil has three kinds of documentation:

• Project documentation, in standard Markdown files.
• Code documentation, inline in code and rendered with Rustdoc.
  • Render with cargo doc or cargo doc --open.
• Algorithm documentation, in Typst files and rendered to PDF.
  • Works out the math behind some of the algorithms implemented in sciutil.
  • Render with just typst-doc or per-file with typst compile.

License

Sciutil is licensed under the Mozilla Public License, version 2.0 or (as the license stipulates) any later version. A copy of the license should be distributed with sciutil, located at LICENSE, or you can obtain one at https://mozilla.org/MPL/2.0/.