File Exchange

image thumbnail


version 1.5 (38.1 KB) by

Dimensioned variables with enforced dimensional consistency



The `physunits` toolbox is an attempt to confer "dimensional awareness" to the
MATLAB environment. The motivation and a suggested algorithm are explained in:
Petty, G.W., 2001. Automated computation and consistency checking of physical
dimensions and units in scientific programs. _Software: Practice and Experience_,
31(11), pp.1067-1076. The author also made available for download a FORTRAN90
module that implements this idea in the FORTRAN language. The module and paper can
be downloaded from The
`physunits` toolbox is based on that module and expands it while trying to adhere
consistently to MATLAB standards and practices (_from 2006_).
The `physunits` toolbox works by defining the `preal` data type and overloaded
functions to support it. A variable of type `preal` represents a physical
quantity. It has two _private_ fields: 'value' - the numerical value, which must
be a numeric type with zero imaginary part, and 'units' - a vector of 7 numbers,
representing the physical dimensions associated with 'value'. The format for the
'units' vector is [length, mass, time, temperature, electric current, amount of
matter, illumination, luminous intensity]. But these details are never needed by
the user, who defines his or her variables via an interface structure. The
interface structure contains predefined variables of type `preal`, representing
the most useful SI units as well as many other units, derived units, constants of
nature, parameters, etc. Get this structure by calling the `setUnits` function.
Create your own dimensioned variables by multiplying a number with an existing
`preal` variable.

See the HTML documentation for more details and examples.

### Setup

The physunits toolbox is implemented as an old-style (pre-2008) MATLAB class. In
this style MATLAB classes are defined by a directory of the class name preceded by
a `@` sign. The `@preal` directory contains the `preal` class definition and
methods. **To use the `preal` class, the directory immediately above `@preal` must
be on the MATLAB search path.**

Comments and Ratings (3)

Very nicely done, thank you! I always wondered why Matlab does bring this feature natively.

Riccardo Meldolesi

this toolbox seems pretty good. The list of overloaded matlab functions is pretty exhaustive.

A wish could be to have the possibility to type directly the unit name rather than access a subfield of the @preal class, for example:

aVariable = 12 * m

instead of:

aVariable = 12 * si.metre

I could suggest a couple ow ways to do it in a OO way.

A small improvement to an alreadyvery good piece of code.



Moving the submission to GitHub for easier maintenance


Adding an overloaded cumsum function.


I modified the < and <= operators to avoid dimension checking in the special case where the second argument has value zero. The only reason is to allow the common usage of
>> validateattributes(d, {'numeric'}, {'positive'})


Added setFUnits.m to create a "fake units" structure with only type double for deployment to users who do not wish to install the toolbox. Deployment usually involves a single line (single character really) change to code.
Update documentation.

Added functions:

Moved to category Utilities:Development environment

MATLAB Release
MATLAB 7 (R14)

Download apps, toolboxes, and other File Exchange content using Add-On Explorer in MATLAB.

» Watch video

Win prizes and improve your MATLAB skills

Play today