The Central Idea
Consider the function
\[ f(x)=2x+1. \]It can be represented in several ways.
As a formula:
\[ f(x)=2x+1. \]As a table:
| \(x\) | \(f(x)\) |
|---|---|
| 0 | 1 |
| 1 | 3 |
| 2 | 5 |
| 3 | 7 |
As an algorithm:
- take the input;
- multiply it by \(2\);
- add \(1\).
As a graph, the same input-output pairs lie on a straight line.
These are not four different functions. They are four representations of the same underlying relationship.
A Function Is Not Its Formula
A formula is one way to describe a function, but a function can exist without a convenient formula.
For example, a table of measured values may define a function on a finite set of inputs even if no simple algebraic expression is known.
Representation Versus Object
The mathematical object is the function itself: an assignment from inputs to outputs.
A representation is a way of showing or communicating that assignment.
This distinction is useful because different representations emphasize different features.
The Four Main Representations
In elementary mathematics, functions are often represented as:
- tables — selected input-output pairs;
- graphs — geometric locations of those pairs;
- formulas — symbolic rules;
- algorithms — procedures for producing outputs.
Tables
A table lists specific inputs and their corresponding outputs.
For example:
| \(x\) | \(f(x)\) |
|---|---|
| \(-2\) | 4 |
| \(-1\) | 1 |
| 0 | 0 |
| 1 | 1 |
| 2 | 4 |
This table is consistent with
\[ f(x)=x^2. \]What Tables Reveal Well
Tables are especially useful for:
- exact listed values;
- experimental or measured data;
- discrete domains;
- repeated outputs;
- patterns across selected inputs;
- comparing several quantities side by side.
What Tables Hide
A finite table usually does not show every input.
It may hide:
- behavior between listed values;
- turning points not sampled;
- domain gaps not shown;
- asymptotic behavior;
- the exact underlying rule.
A Table Can Be the Whole Function
If the domain is finite, a complete table may fully define the function.
For example,
\[ f:\{1,2,3\}\to\{4,7,9\} \]can be completely represented by three listed assignments.
A Table Can Also Be Only a Sample
If the function has a continuous domain, a table usually shows only selected values.
For example, four sampled values from
\[ f(x)=\sin x \]do not reveal the entire continuous wave.
Graphs
A graph represents input-output pairs as points.
For a function
\[ y=f(x), \]each input-output pair
\[ (x,f(x)) \]becomes a point in the coordinate plane.
What Graphs Reveal Well
Graphs are especially useful for seeing:
- increasing and decreasing behavior;
- maximum and minimum values;
- zeros and intercepts;
- symmetry;
- continuity and jumps;
- long-term trends;
- turning points;
- relative rates of change;
- overall shape.
What Graphs Can Hide
A graph may not show exact values clearly.
Its appearance also depends on:
- window size;
- axis scale;
- sampling density;
- plotting accuracy;
- whether important features lie off-screen.
Graphing Window Matters
The same function can look very different under different axis scales.
A steep curve can appear nearly flat if the vertical scale is compressed.
A small oscillation can appear dramatic if the vertical scale is magnified.
Graphs must be read with their axes.
Graphs Show Structure, Not Just Points
One major advantage of graphs is that many relationships become visually apparent at once.
A table may require careful comparison row by row, while a graph can reveal a turning point or trend immediately.
Formulas
A formula gives a symbolic rule connecting inputs and outputs.
For example,
\[ f(x)=3x^2-2x+5. \]The formula can generate infinitely many outputs without listing them individually.
What Formulas Reveal Well
Formulas are especially useful for:
- exact symbolic structure;
- algebraic manipulation;
- solving equations;
- finding domains;
- identifying parameters;
- generalizing patterns;
- proving properties.
What Formulas Can Hide
A formula may make the overall behavior difficult to see immediately.
For example,
\[ x^4-4x^3+2x^2+8x-3 \]contains exact information, but its graph shape is not obvious at a glance.
Equivalent Formulas Can Reveal Different Features
Consider
\[ f(x)=x^2-6x+5. \]Factored form:
\[ f(x)=(x-1)(x-5) \]reveals zeros.
Vertex form:
\[ f(x)=(x-3)^2-4 \]reveals the minimum.
Expanded form emphasizes coefficients.
Representation Within Representation
Even within symbolic formulas, several equivalent forms may function like different representations because each exposes different structure.
Choosing a useful form is part of mathematical problem solving.
Algorithms
An algorithm is a procedure for producing an output from an input.
For
\[ f(x)=2x+5, \]one algorithm is:
- multiply the input by \(2\);
- add \(5\);
- return the result.
Algorithms Can Define Functions
A function does not require a closed formula if there is an unambiguous procedure that produces exactly one output for each allowed input.
Algorithm Example Without a Simple Formula
Define \(f(n)\) for a positive integer \(n\) as:
- write \(n\) in decimal notation;
- count the number of digits;
- return that count.
This algorithm defines a function from positive integers to positive integers.
Algorithms Can Include Decisions
An algorithm may contain conditional steps.
For example:
- if \(x\ge0\), return \(x\);
- if \(x<0\), return \(-x\).
This algorithm defines
\[ f(x)=|x|. \]Piecewise Functions and Algorithms
Piecewise notation and conditional algorithms often describe the same structure.
For example,
\[ f(x)= \begin{cases} -x, & x<0,\\ x, & x\ge0 \end{cases} \]is the symbolic version of a decision procedure.
Algorithms Can Be Iterative
An algorithm may repeat steps.
For example, a numerical method may repeatedly update
\[ x_{n+1}=f(x_n) \]until a stopping condition is satisfied.
Such procedures connect functions to computation and numerical analysis.
Algorithms Need to Be Unambiguous
A valid function-producing algorithm must specify exactly one output for every allowed input.
If a procedure leaves choices unresolved, it may fail to define a function.
The Same Function in Four Forms
Consider
\[ f(x)=x^2+1. \]Formula:
\[ f(x)=x^2+1. \]Algorithm:
- square the input;
- add \(1\).
Table:
| \(x\) | \(f(x)\) |
|---|---|
| \(-2\) | 5 |
| \(-1\) | 2 |
| 0 | 1 |
| 1 | 2 |
| 2 | 5 |
Graph: an upward-opening parabola with minimum value \(1\).
Translate Formula to Table
Suppose
\[ f(x)=3x-2. \]Choose inputs and evaluate:
| \(x\) | \(f(x)\) |
|---|---|
| 0 | \(-2\) |
| 1 | 1 |
| 2 | 4 |
This is a direct translation from symbolic to numerical representation.
Translate Table to Graph
Each row
\[ (x,f(x)) \]becomes one point on the graph.
For example, the row
\[ x=2,\qquad f(x)=4 \]becomes the point
\[ (2,4). \]Do Not Automatically Connect Table Points
If the domain is discrete, connecting the points may falsely suggest outputs exist between them.
For example, a function whose input is the number of employees may have integer-only inputs.
The graph should respect the domain.
Continuous and Discrete Graphing
For a continuous real interval domain, a curve may connect the points.
For a discrete domain, isolated points are usually more accurate.
Representation should preserve the nature of the domain.
Translate Graph to Table
A graph can be sampled at selected input values to create a table.
But the resulting table usually contains less information than the full graph because only chosen points are retained.
Translation Can Lose Information
Moving from one representation to another may discard information.
For example:
- a sampled table may lose behavior between points;
- an equation obtained from a parametrization may lose orientation;
- a graph may hide exact symbolic values;
- a simplified formula may hide original domain restrictions.
Translate Table to Formula
Sometimes a table suggests a formula.
Consider:
| \(x\) | \(y\) |
|---|---|
| 1 | 5 |
| 2 | 8 |
| 3 | 11 |
| 4 | 14 |
The output increases by \(3\) whenever \(x\) increases by \(1\).
This suggests a linear rule with slope \(3\).
Checking one point gives
\[ \boxed{y=3x+2}. \]A Pattern Does Not Always Determine a Unique Formula
A finite table does not generally determine one unique formula.
Many different functions can agree on the listed points and differ elsewhere.
Extra assumptions are needed, such as:
- the function is linear;
- the function is quadratic;
- the simplest pattern is intended;
- a particular model family has been chosen.
Interpolation
Finding a function that passes through known data points is called interpolation.
There may be many possible interpolating functions unless the function family is restricted.
Model Fitting Is Different from Exact Interpolation
Real data may contain noise or measurement error.
Instead of forcing a formula through every point, we may choose a model that approximately follows the overall pattern.
This leads to regression and statistical modeling.
Translate Formula to Algorithm
Consider
\[ f(x)=\frac{x^2+1}{3}. \]An algorithm is:
- square the input;
- add \(1\);
- divide by \(3\).
Translate Algorithm to Formula
Suppose the procedure says:
- subtract \(4\) from the input;
- square the result;
- add \(7\).
Then the formula is
\[ \boxed{f(x)=(x-4)^2+7}. \]Order Matters in Algorithms
Compare:
- subtract \(4\);
- square.
with:
- square;
- subtract \(4\).
The first gives
\[ (x-4)^2, \]while the second gives
\[ x^2-4. \]Procedural order and function composition are closely connected.
Algorithms and Function Composition
A multistep algorithm can often be understood as a composition of simpler functions.
For example:
\[ x \mapsto x-4 \mapsto (x-4)^2 \mapsto (x-4)^2+7. \]Graphs and Algorithms
An algorithm can generate points that are then plotted.
This is how computers graph many functions: they evaluate the rule at many inputs and display the resulting coordinates.
Computer Graphs Are Sampled Representations
A digital graph is usually created from a finite number of evaluations.
The software then draws pixels or line segments between samples.
This can occasionally create misleading visual behavior if the sampling is too coarse.
Numerical Representation
Tables are one form of numerical representation, but numerical descriptions can also include:
- lists;
- arrays;
- data files;
- spreadsheets;
- computed samples;
- lookup tables.
Lookup Tables
Some functions are implemented by storing precomputed values rather than evaluating a formula each time.
This is common when:
- evaluation is expensive;
- data come from measurement;
- only certain inputs are needed.
Algorithms Can Use Lookup Tables
A procedure may combine stored data with calculation.
For example:
- find the nearest stored input values;
- retrieve their outputs;
- interpolate between them.
This is a hybrid representation.
Functions in Programming
In programming, a function often looks like an algorithm that accepts inputs and returns outputs.
This resembles the mathematical idea closely, though programming functions may also have side effects or internal state.
A pure mathematical function is determined entirely by its input and fixed rule.
Deterministic Algorithms
A deterministic algorithm always gives the same output for the same input under the same fixed conditions.
Such an algorithm naturally represents a mathematical function.
Randomized Procedures Need Care
A procedure involving random choices may produce different outputs from the same input.
That procedure is not an ordinary deterministic function from input to output unless the random seed or random outcome is included as part of the input.
Verbal Rules
A function can also be described in ordinary language.
For example:
“Assign each U.S. state its capital city.”
This defines a function even though it is not written as a numerical formula.
Diagrams and Mappings
Arrow diagrams are another representation of functions.
They make domain, codomain, range, injectivity, and surjectivity especially easy to see for finite sets.
No Representation Is Always Best
The best representation depends on the task.
For example:
- use a table to inspect exact sampled data;
- use a graph to understand overall behavior;
- use a formula for symbolic manipulation;
- use an algorithm to describe computation or procedure.
Use Multiple Representations Together
Strong mathematical understanding often comes from combining representations.
A formula may reveal exact structure, a graph may reveal behavior, and a table may confirm selected values.
Each can check and clarify the others.
Representation as a Checking Strategy
Suppose an algebraic solution predicts a zero at
\[ x=4. \]You can:
- substitute \(4\) into the formula;
- check whether the table gives output \(0\);
- look for an \(x\)-intercept near \(4\) on the graph.
Agreement among representations increases confidence in the result.
Representation Can Reveal Errors
If a formula predicts a rapidly increasing function but the graph is falling, something is inconsistent.
Cross-representation checks are powerful because different errors often appear differently in different forms.
Units Must Be Preserved Across Representations
If a function maps hours to miles, the table headings, graph axes, formula interpretation, and algorithm should all preserve those units.
A representation that loses units can become mathematically ambiguous.
Domain Must Be Preserved Across Representations
Suppose
\[ f(n)=2n+1 \]for positive integers \(n\).
If we draw a continuous line through the plotted points without noting the discrete domain, we have changed the represented function.
Codomain May Not Be Visible
A formula or graph often reveals the range but may not reveal the declared codomain.
Surjectivity therefore cannot always be determined from a graph or formula alone.
Piecewise Representations
A piecewise function may be represented:
- symbolically with cases;
- graphically with separate branches;
- numerically with rows from different regions;
- algorithmically with conditional statements.
Parametric Representations
A parametric curve uses functions such as
\[ x=x(t), \qquad y=y(t). \]A table may list
\[ t,\quad x(t),\quad y(t), \]while a graph displays only the resulting coordinate points.
The graph may lose information about timing or orientation.
Implicit Representations
An implicit equation such as
\[ x^2+y^2=25 \]describes a relation without isolating one variable.
A graph shows the whole circle, while explicit formulas require separate branches.
Different representations may therefore organize the same relation differently.
Representation and Function Families
A parameterized formula such as
\[ f(x)=a(x-h)^2+k \]represents an entire family.
A graph may show several members at once, while a table may compare outputs for selected parameter values.
Worked Example: One Relationship, Four Representations
Suppose a machine produces items at a constant rate of \(8\) items per hour.
Formula:
\[ P(t)=8t. \]Algorithm:
Multiply hours worked by \(8\).
Table:
| Hours \(t\) | Items \(P(t)\) |
|---|---|
| 1 | 8 |
| 2 | 16 |
| 3 | 24 |
Graph:
A line through the origin with slope \(8\).
Worked Example: Table to Formula
| \(x\) | \(f(x)\) |
|---|---|
| 0 | 4 |
| 1 | 7 |
| 2 | 10 |
| 3 | 13 |
The outputs increase by \(3\) for each increase of \(1\) in the input.
So the slope is
\[ 3. \]Since
\[ f(0)=4, \]the formula is
\[ \boxed{f(x)=3x+4}. \]Worked Example: Algorithm to Formula
Procedure:
- add \(2\) to the input;
- square the result;
- subtract \(5\).
The corresponding formula is
\[ \boxed{f(x)=(x+2)^2-5}. \]Worked Example: Formula to Algorithm
Let
\[ f(x)=\sqrt{3x+1}. \]Algorithm:
- multiply the input by \(3\);
- add \(1\);
- take the principal square root.
The real domain also requires
\[ 3x+1\ge0. \]A Reliable Representation Workflow
- Identify the function's domain and output meaning.
- Choose the representation that best answers the current question.
- Preserve units, restrictions, and discrete-versus-continuous structure.
- Translate carefully when changing representation.
- Ask what information may be lost in the translation.
- Use a second representation to check important conclusions.
A Reliable Table-to-Formula Workflow
- Order the inputs.
- Look for constant differences, constant ratios, or other recognizable patterns.
- Choose a plausible function family.
- Determine the parameters from the data.
- Verify the proposed formula against all listed pairs.
- Remember that finite data rarely determine a unique formula without assumptions.
A Reliable Formula-to-Graph Workflow
- Identify the function family.
- Determine the domain.
- Find useful landmarks such as intercepts, extrema, asymptotes, or known points.
- Determine increasing, decreasing, or constant behavior.
- Choose an axis scale that reveals the important features.
- Check the graph against direct evaluations.
Common Errors
Assuming a Function Must Have a Formula
Tables, mappings, verbal rules, and algorithms can define functions too.
Assuming a Finite Table Determines One Unique Formula
Many different functions can agree on finitely many points.
Connecting Discrete Data as if the Domain Were Continuous
This may invent outputs that are not part of the function.
Ignoring Axis Scale
A graph's visual appearance depends strongly on the chosen window and scale.
Assuming a Graph Gives Exact Values
Graphs are often approximate visual representations.
Forgetting Domain Restrictions During Translation
An algebraically simplified formula may hide exclusions present in the original definition.
Ignoring Units
Units belong in tables, graph labels, formulas, and algorithms when the function models real quantities.
Using Only One Representation
A second representation often reveals structure or errors that the first hides.
Practice
1. Formula to Table
Let
\[ f(x)=2x-3. \]Find \(f(0)\), \(f(1)\), and \(f(2)\).
Check Answer
\[ f(0)=-3, \] \[ f(1)=-1, \] \[ f(2)=1. \]2. Algorithm to Formula
Write a formula for this procedure:
- multiply the input by \(4\);
- subtract \(7\).
Check Answer
\[ \boxed{f(x)=4x-7}. \]3. Formula to Algorithm
Describe an algorithm for
\[ f(x)=(x-2)^2+6. \]Check Answer
- subtract \(2\) from the input;
- square the result;
- add \(6\).
4. Table Pattern
| \(x\) | \(y\) |
|---|---|
| 1 | 6 |
| 2 | 10 |
| 3 | 14 |
Assuming a linear relationship, find the formula.
Check Answer
The slope is \(4\).
Using \((1,6)\):
\[ 6=4(1)+b, \] \[ b=2. \]So
\[ \boxed{y=4x+2}. \]5. Discrete or Continuous?
A function gives the number of possible committees that can be chosen from \(n\) people. Should its natural input domain be continuous or discrete?
Check Answer
Discrete. The input \(n\) counts people and is typically a nonnegative integer.
6. Best Representation
Which representation is especially useful for seeing turning points and overall shape?
Check Answer
A graph.
7. Exact Symbolic Structure
Which representation is usually strongest for algebraic manipulation and proof?
Check Answer
A formula or symbolic expression.
8. Procedure
Which representation is especially natural for describing how a computer calculates an output step by step?
Check Answer
An algorithm.
Key Ideas
- A function is a relationship, not a particular notation.
- The same function can have numerical, graphical, symbolic, and procedural representations.
- Tables show exact listed values and are especially useful for discrete or measured data.
- Graphs reveal global behavior, shape, extrema, intercepts, and trends.
- Formulas reveal symbolic structure and support algebraic manipulation.
- Algorithms describe how outputs are produced step by step.
- A function does not need a simple closed formula to exist.
- Finite tables generally do not determine a unique formula without additional assumptions.
- Different equivalent formulas can reveal different structural features.
- Translation among representations can lose information.
- Domain, codomain, units, and discrete-versus-continuous structure must be preserved.
- Multiple representations are often most powerful when used together.
- Algorithms connect function theory to computation, while graphs connect it to geometry and visualization.
Where This Leads
Moving fluently among tables, graphs, formulas, and algorithms makes functions usable across mathematics, science, engineering, data analysis, and computation.
These representations lead naturally into numerical modeling, interpolation, regression, graphing technology, algorithms, recurrence relations, simulation, data visualization, symbolic algebra, and the broader practice of choosing the representation that best exposes the structure needed for a problem.