simple_decimal - Precise Decimal Arithmetic for Eiffel

Overview

simple_decimal is a clean wrapper over Gobo's MA_DECIMAL library that provides arbitrary-precision decimal arithmetic for Eiffel. Perfect for financial calculations where binary floating-point errors are unacceptable.

The Problem with Floating-Point

Binary floating-point (REAL, DOUBLE) cannot represent many decimal fractions exactly:

-- DOUBLE fails silently!
io.put_double (0.1 + 0.2)  -- Outputs: 0.30000000000000004  WRONG!

This causes real-world problems: financial calculations that don't balance, prices that display incorrectly, tax calculations off by a penny, and cumulative errors in accounting systems.

The Solution

local
    a, b, total: SIMPLE_DECIMAL
do
    create a.make ("0.1")
    create b.make ("0.2")
    total := a + b
    print (total.out)  -- Outputs: 0.3  CORRECT!
end

Quick Start

Installation

Set the environment variable:
```
export SIMPLE_EIFFEL=/d/prod
```

Add to your ECF file:

<library name="simple_decimal" location="$SIMPLE_EIFFEL/simple_decimal/simple_decimal.ecf"/>

Basic Usage

class
    INVOICE_CALCULATOR

feature -- Calculation

    calculate_total (subtotal_str, tax_rate_str: STRING): STRING
            -- Calculate total with tax, returning formatted currency
        local
            subtotal, tax_rate, tax, total: SIMPLE_DECIMAL
        do
            create subtotal.make (subtotal_str)
            create tax_rate.make (tax_rate_str)

            -- Calculate tax: subtotal * (rate / 100)
            tax := (subtotal * tax_rate.from_percentage).round_cents
            total := subtotal + tax

            Result := total.to_currency_string
        end

end

Result: calculate_total ("47.99", "8.25") returns "$51.95"

Key Features

Arbitrary Precision

No floating-point representation errors. 0.1 + 0.2 = 0.3 exactly.

Financial Operations

Currency formatting, percentages, tax calculations, bill splitting with penny-perfect accuracy.

Multiple Rounding Modes

Banker's rounding (half-even), round up, round down, ceiling, floor, truncate.

Immutable Operations

All arithmetic returns new values. Thread-safe, predictable, no side effects.

Smart Parsing

Handles "$1,234.56", "1234.56", "-$99.99" automatically.

Design by Contract

Full preconditions, postconditions, and class invariants throughout.

Documentation

User Guide - Complete tutorial with examples
API Reference - Full feature documentation
Architecture - Internal design and rationale
Cookbook - Real-world recipes and patterns

Standards Compliance

simple_decimal is built on Gobo's MA_DECIMAL which implements:

General Decimal Arithmetic Specification - IBM's decimal arithmetic standard
IEEE 754-2008 - Standard for floating-point arithmetic (decimal64/decimal128)

This means your calculations follow the same rules used by financial institutions worldwide.