Lorem Ipsum Generator

How to use the lorem ipsum generator

Choose a unit, choose a count, decide whether to open with the canonical phrase, and the text appears immediately. The defaults give three paragraphs beginning Lorem ipsum dolor sit amet, consectetur adipiscing elit, which comes to 156 words and 1,059 characters across 13 sentences.

Sentences run between 8 and 16 words and paragraphs between 3 and 5 sentences, both varied within those bounds so the block has an uneven right edge and an uneven depth, exactly as real copy does. A comma is placed roughly two thirds of the way through any sentence longer than ten words, where a real clause would break.

Placeholder text has one job: to let you judge a layout without reading it. Real copy in a mockup is a trap, because everyone in the review starts editing the sentences instead of looking at the line length, the leading, the measure, and what happens when a paragraph runs two lines longer than the box allows.

Nonsense Latin removes that temptation while keeping the statistical shape of Western prose: word lengths in a familiar distribution, ascenders and descenders in familiar proportions, and a texture on the page that behaves like text rather than like a grey rectangle.

The generator is seeded and deterministic, which is an engineering requirement rather than a stylistic choice. This page renders on the server and then hydrates in the browser, and the two renders must produce identical markup, so the words cannot be left to an unseeded random call.

The same property makes the tool useful elsewhere: identical settings give identical text, so a test fixture or a screenshot comparison built on it stays stable between runs rather than changing underneath the thing it was meant to hold still.

Where the Latin comes from

The words come from a 119-word pool, 115 of them distinct, drawn from the standard lorem ipsum boilerplate and the extended passages that usually accompany it. That vocabulary descends from De Finibus Bonorum et Malorum, a work on ethics that Cicero wrote in 45 BC.

The connection was traced by Richard McClintock, a Latin scholar at Hampden-Sydney College, who followed the unusual word consectetur back to sections 1.10.32 and 1.10.33 of that text. The passage itself is Cicero setting out the argument that nobody desires pain for its own sake, which is why the fragment reads dolorem ipsum quia dolor sit amet.

The standard passage is a corrupted extract of that text, with words truncated, reordered, and letters altered until the Latin no longer parses. Typesetters had been using the scrambled version as specimen text since at least the sixteenth century, and it reached the modern industry through Letraset transfer sheets in the 1960s and desktop publishing software in the 1980s, which is how a Renaissance printing convenience became the default filler text of the web.

The formula, worked line by line

There is no arithmetic to speak of. What matters is the shape of the generated text and the fact that the same inputs always produce the same output, so the interesting part of the design is the random number generator, which is deliberately the weakest one in this repository.

Word selection runs on xorshift32, a three-line pseudo-random generator seeded with an integer that defaults to 1. It is small, fast, and entirely predictable, which for cryptography would be disqualifying and for placeholder prose is precisely what is wanted: predictable means the server and the browser agree.

state       = seed, default 1
next()      = xorshift32: state ^= state << 13, >>> 17, << 5
word        = pool[floor(next() x 119)]        pool = 119 words, 115 distinct
sentence    = 8 + floor(next() x 9) words, comma at 60 percent if longer than 10
paragraph   = 3 + floor(next() x 3) sentences

The determinism is not decoration. A page like this one is rendered to HTML on the server and then hydrated by the browser, and the two renders must produce identical markup, which a seeded xorshift32 guarantees: same seed, same state transitions, same words.

The predictability that would disqualify this generator for cryptography is precisely the property wanted here, and it is why the strongest and weakest random number generators in this repository sit a few files apart, each correct for its job.

Questions people ask