MTH229 with Julia
Preface
Introduction
MTH229 at the College of Staten Island is a course to introduce a programming language to reinforce concepts of a first-semester calculus course from numerical and graphical points of view.
Some sections use the Julia programming language. For Julia
, the computer-lab “projects” are on WeBWorK and there should be sufficient background material therein to work through the details. In the notes here can be found additional detail for those seeking it. For a more thorough introduction, visit Calculus with Julia.
Installation details for Julia
are at the end of this page, but most students will use a departmental web server to access Julia
. In a pinch, there are online instances (resource-constrained) that can be used:
- Colab. Google’s
colab
service allows the free user more computing power than doesbinder
but does require more setup. This is working as of March 2025:
Go to google colab:
https://colab.research.google.com/
Click on “Runtime” menu and then “Change Runtime Type”
Select Julia as the “Runtime Type” then save
Copy and paste then run this set of commands
using Pkg
Pkg.add("Plots")
Pkg.add("MTH229")
using MTH229
using Plots
This may take 2-3 minutes to load. The plotly()
backend doesn’t work out of the box. Use gr()
to recover if that command is issued.
- Binder. Binder is a free service for small images though it can be very tempermental. Binder has an inactivity time out of a few minutes. Two images are available, the one without
SymPy
loads much faster and would typically suffice, though has no symbolic capabilities.
These notes are broken into different sections, where most all sections have some self-grading questions at the end that allow you to test your knowledge of that material. The code should be copy-and-pasteable into a Julia
session. The code output is similar to what would be shown if evaluated in an IJulia
cell, our recommended interface while learning Julia
, though some may like the Pluto
interface as well.
The notes mostly follow topics of a standard first-semester calculus course after some background material is presented for learning Julia
within a mathematical framework.
Each topic has a lab project. At CSI, sufficient time is allotted to complete these projects during the lab class. These projects are available as IJulia
notebooks or Pluto
notebooks. (There are also Pluto
notebooks with 5-10 minutes of commentary.) Find links for these near the top of each page of notes.
For example, blank notebooks for test taking, etc. are found by following these links:
Lab notebook
A laboratory notebook is a useful tool for keeping notes. One specific for this class is available for printing out: LabNotebook
WeBWorK
Most background material, and the questions are now presented and completed through WeBWorK
.
The provided .ipynb
notebooks only contain modest background details and a few examples.
Each WeBWorK
page has an introduction to some part of the material and a series of questions and comments related to the introduction.
There are a few idiosyncrasies in the WeBWorK
pages to be aware of:
Typesetting of code examples
The code examples may be typeset in WeBWorK
as though they appear in a terminal. The terminal formatting (or REPL, read-evaluate-print-loop) is indicated with a leading prompt julia>
. Interacting with terminal is interactive: a command is read in, evaluated, its output is printed. The output displayed directly below the command which is prefixed with the prompt.
This is in contrast to some examples where a block of code is typeset as though it were the input for a notebook cell. Notebooks take a block of input, evaluate the block, and then print out the last command that was executed. (The use of @show
or print(...)
can be used to display intermediate values in a cell.) Notebooks are how you are expected to interact with Julia
.
Comments in code
The pound sign (#
) in Julia
instructs the interpreter that everything afterwards on the line is a comment. While the Julia
interpreter can ignore what comes after the pound sign, it is expected that the human interpreter gets some additional bit of detail about why the command is as it is.
Typesetting of math
WeBWorK
allows the use of MathJax, a JavaScript display engine, for mathematics giving math a distinct look. In contrast, code is typeset using a typewriter style, fixed width font.
Copy and paste from a
WeBWorK
page
Copying and pasting from WeBWorK
into a notebook will usually be unsuccessful as the interesting parts are typeset with MathJax. For the intrepid, the rendered text can be selected by using “right-click” (or “control-click” on a mac) and the “Copy to Clipboard” with “TeX Commands”. Some editing will be necessary.
Numeric answers
For most questions with a numeric answer it is best to copy all 16 digits of output. Several digits are expected to match a correct answer. For numeric questions where an estimate is made, say from a graph, this is significantly relaxed.
Scientific notation
While Julia
is very happy to express its output using scientific notation, WeBWorK
is not happy to receive the exact output for an answer. Either replace e
with E
(as in 1.23e4
would be 1.23E4
) or use decimals.
Lists of numbers
Many questions expect one or more answers. Separating each one with a comma is needed for WeBWorK
to distinguish the different answers.
Other answers
If the answer is to be a function, the automatic grader is expecting just the rule of the function (an expression), as in for f(x) = sin(x)
just enter sin(x)
.
Basic ideas
Julia
makes an excellent choice as a computer supplement for this material as its syntax is very similar to standard mathematical syntax. The ability to define mathematical functions using the familiar f(x) = ...
notation makes getting started really easy. Further, the fact that functions are first-class objects means that it is possible to create higher-order Julia
functions that mirror the standard operations of calculus. The following pattern is used throughout:
action(function_object, args...)
For example, the notes use:
plot(f, a, b)
to plotf
over[a,b]
;plot!(g)
to add a curve (usingPlots
)find_zero(f, (a, b))
orfzero(f, a, b)
to find a zero between the bracketing interval[a,b]
(fromRoots
)find_zero(f, a)
(orfzero(f, a)
) to find a zero near an initial starting pointa
(fromRoots
)lim(f, c)
to numerically investigate the limit off
atx=c
.limit(f(x), x=>c)
to find the limit off
atc
(fromSymPy
)f'
to return a function that computes the derivative off
(Added in theMTH229
package based on thederivative
function from theForwardDiff
package)diff(f(x),x)
to find a symbolic derivative off
(fromSymPy
)find_zeros(f, (a, b))
(orfzeros(f, a,b)
) to find all the real zeros of a functionf
in[a,b]
(fromRoots
)sign_chart(f, a, b)
to find and classify zeros of a function over[a,b]
quadgk(f, a, b)
to find the numeric integral off
over(a,b)
(from theQuadGK
package)integrate(f(x), x)
to find the symbolic integral off
(from theSymPy
package)integrate(f(x), (x, a, b))
to find the definite integral over[a,b]
symbolically (from theSymPy
package).
With just this basic set of actions, akin to buttons on the calculator, a rich variety of problems can be addressed.
Help and errors
Most functions in Julia
are documented. To access the documentation we issue the command prefixed with a question mark, as in ?sin
. At the REPL, typing a leading ?
turns on the help mode. In Jupyter, the query should be the only thing in the cell.
Help can be narrowed to respond to just certain argument types.
Errors
There are a variety of common errors encountered when first interacting with Julia
. We try to enumerate them here.
Parsing errors
ERROR: MethodError: no method matching ^(::typeof(sin), ::Int64)
: This error is thrown when trying to executesin^2(3)
. The issue is there is no means to square the function object, rather we square the value of the function after calling it on3
:sin(3)^2
ERROR: ParseError:...space required before '?' operator
This came when trying to use the ternary operator without spaces around the?
and:
(by typingx<1?2:3
and notx<1 ? 2 : 3
).ERROR: ParseError: whitespace is not allowed here
The error came by including a space after the function name and before the opening parentheses when calling the function. That issin (3)
errors,sin(3)
does not.ParseError: Expected ')'
This came because a non-matching number of parentheses was used – fewer closing than opening. The explanatory message will beextra tokens after end of expression
if there are more closing than opening.
Incorrect function usage
ERROR: DomainError with -1.0:
This happened when putting a negative (real) number into thelog
function, in particular enteringlog(-1)
. A similar error will happen withsqrt(-1)
and withsin(Inf)
, though each has particular clarifying language.ERROR: MethodError: objects of type Float64 are not callable
This occurs when there is a missing multiplication sign. For example the expression(sin(3))(sin(4))
will error, as it sees the parentheses aroundsin(4)
as a function call. However, such notation does work with numeric literals, so it can be confusing. We recommend using*
when you mean multiply.
Technical errors
ERROR: cannot define function h; it already has a value
.Julia
has two types of functions anonymous and generic. Generic functions are what we mostly model in this class. They are called generic as each function can have many different implementations (methods). To make this work, a lookup-table is used to store the function name and the number and type of arguments the method is defined for. To do this, once a generic function is defined, it can’t be repurposed as a different variable. Similarly, if you use a name as a variable (eg.h=2
) it can no longer be used as a generic function name (eg.h(x) = sin(x)
). We recommend sticking to familiar math names for variables (e.g.x
,y
,a
, …) and familiar math names for functions (e.g.,f
,g
, …). This works pretty well, but sometimes there is crossover, such as withh
.
Oops and not an error
- It is possible to redefine and shadow basic math operations. For example, if you were to type the following equation into
Julia
:
x + y = 3
you will not get an equation, rather this will redefine the addition of two variables, always returning 3
. Oops. Often redefinition causes and error, such entering this equation
sin(x) = 3 # ERROR: invalid method definition in Main: function Base.sin must be explicitly imported to be extended
To break out of this is tough, you need to redefine +
, reaching into Base.:+
to do so
x + y = Base.:+(x,y)
We recommend learning fast that equations are not entered directly into Julia
.
Installing Julia
on a personal laptop or computer.
Installing Julia
is not terribly difficult, but does involve a few additional steps. Use juliaup
to manage a Julia
installation
On Windows from the command line run
winget install julia -s msstore
On Mac or Linux, from the command line run
curl -fsSL https://install.julialang.org | sh
These two command install juliaup
, which you then run from the command line to install Julia
.
Details on additional packages to install for MTH229 are listed at https://github.com/mth229/MTH229.jl/.
This table covers pros and cons for the approaches mentioned above:
Using server Colab Binder Local Installation
Setup ease ✓ ~ ✓ ×
Speed ✓ ✓ × ✓
Persistence of work × ✓ × ✓
Free ✓ ✓ ✓ ✓
Use at home ✓ ✓ ✓ ✓