Scientific Calculator

Using the Scientific Calculator

1. Click the on-screen buttons or type directly with your keyboard. Digits, the four arithmetic operators, parentheses, and the caret for powers all work.
2. Call a function by tapping its named key - sin, cos, tan, log, ln, or √ - then the argument, then the closing parenthesis. The display previews the expression as you build it.
3. Insert a constant with the π or e key. They expand to the full double-precision value, not a truncated 3.14 or 2.71.
4. Press = or Enter to evaluate. The expression plus its answer is pushed to the history strip above; the last five calculations stay visible for reuse.
5. Edit with Backspace (the left-arrow key on the keypad) to remove the last character, or C to clear the entire expression. Escape also clears.

What Happens When You Press Equals

The calculator implements a small recursive-descent parser in TypeScript rather than delegating to JavaScript's eval. Tokens are scanned left-to-right into numbers, operators, function names, and parentheses; a shunting-yard-style precedence table then lowers the token stream into an AST of binary and unary nodes before evaluation. This matters for security (the display can never execute arbitrary JS) and for correctness (the parser enforces ISO 80000-2 operator precedence rather than inheriting JavaScript's quirks like 0.1 + 0.2 !== 0.3, which you will still see in the output because it is a property of IEEE 754 binary64 arithmetic).

Trig, log, and root functions call straight through to Math.sin, Math.cos, Math.log10, Math.log, and Math.sqrt, so you get whatever your engine uses: V8 (Chrome, Edge), SpiderMonkey (Firefox), or JavaScriptCore (Safari). All three implement IEEE 754 elementary functions to roughly one ULP, which is why sin(π) prints as a tiny number near 10^-16 instead of a clean zero.

When a Scientific Calculator Is the Right Choice

• Checking homework on trigonometric identities or quickly evaluating tan(π/4).
• Converting decibels or pH: log(x) gets you base-10, ln(x)/ln(10) if you forget there is a dedicated key.
• Computing compound interest with (1 + r)^n without firing up a spreadsheet.
• Sanity-checking a numeric result from a longer calculation in Python, Excel, or Wolfram.
• Prototyping on a phone or tablet where the OS calculator refuses to rotate or lacks trig functions.
• Teaching order of operations to a student, with the history strip showing exactly how expressions parse.

Gotchas and Floating-Point Precision

• Radians, not degrees. sin(90) returns 0.893..., not 1. Convert with angle * π / 180 or type sin(π/2) directly.
• Binary floats can't represent 0.1 exactly. 0.1 + 0.2 evaluates to 0.30000000000000004 in every IEEE 754 engine. This is arithmetic reality, not a bug.
• Subtractive cancellation. 1 + 1e-16 - 1 returns 0, because the intermediate result rounds to 1 at double precision. Reorder your expression to avoid cancelling a big number against a tiny one.
• Division by zero returns Infinity, not an error, following the IEEE 754 spec. 0/0 returns NaN. Both render as readable strings.
• Implicit multiplication is not supported. Write 2*(3+4), not 2(3+4); the latter is treated as a call to a non-existent function.
• Negative powers of negative bases are tricky: (-2)^0.5 is NaN because the result is complex, while (-2)^2 is 4. The tool does not do complex arithmetic.

A Brief History of Calculator Notation

What the calculator implements is the infix algebraic notation standardized by Hewlett-Packard's competitor Texas Instruments in the SR-50 (1974), and formalized in ISO 80000-2 for mathematical signs and symbols. That standard defines precedence (unary minus and exponentiation above multiplication, multiplication above addition, left-to-right within a tier), the meaning of log as base-10 outside pure mathematics, and the convention that sin applied without an argument in parentheses is malformed. HP's original 1968 machines used Reverse Polish Notation (RPN) instead, which needs no parentheses at all; RPN survives in stack-based languages and some trader keyboards but lost the consumer war.

Browser Calculator vs. The Alternatives

A physical calculator (TI-30, Casio fx-991) runs its evaluator on fixed-point BCD hardware and can be more accurate for long chains of division; a scientific calculator app like PCalc offers programmable memory and unit conversion; a desktop REPL (Python with math, Julia, or the Windows Calculator in Programmer mode) is better for anything involving arrays, symbolic algebra, or high precision. Wolfram Alpha will solve equations; this tool will not. Where the browser version wins is latency and availability: press "/" in your URL bar, compute, and close the tab, without installing anything or signing in. For quick arithmetic and trig with the occasional log and square root, it replaces the OS calculator one to one.