Visual net spins literary web

Paley: Software can be used for text analysis

By Jonathan Fildes BBC Science

A website that puts a new spin on thousands of literary classics has been created by an American designer.

Textarc.org is an interactive program that reproduces the text of more than 2,000 books as works of art.

The software converts the text into an interactive map that allows viewers to quickly see relationships between words and characters at a glance, even without having read the book.

"It is a piece of art" said the founder of the site, W Bradford Paley, a designer for Digital Image Design Incorporated in New York.

"But I originally built it to be a text analysis tool," he told the BBC programme Go Digital.

Word spirals

The texts, which range from Shakespearean classics like Hamlet to Lewis Carroll's Alice's Adventures in Wonderland, are supplied by Project Gutenberg, a free online library.

It's as if I wrote the word once and attached it by rubber bands to the places it should be in the story

W Bradford Paley, TextArc

Once a book has been chosen, the program reprints the text line by line, in a wide oval around the screen.

The text is almost unreadable but shows the viewer the typographic structure of the book. A second, concentric ellipse then appears inside the first with the text in a more readable font.

Inside the spirals more words appear. Words that occur more than once are distributed inside the ellipses in their average position.

The brighter the word the more frequently it occurs and the more important it is to the story.

'Rubber bands'

Rolling your mouse over a word creates lines that link it to every occurrence in the elliptical text.

The most used words are brighter

For example Alice glows at the centre of a map of Carroll's work. She is attached to the inner ellipse by hundreds of lines, showing the viewer that she appears throughout the book.

In contrast, scrolling over Gryphon or Mock Turtle shows fewer links to just the chapters where they appear, and their position on the periphery shows they do not occur through out the book.

"It's as if I wrote the word once and attached it by rubber bands to the places it should be in the story" explained Mr Paley.

"Alice is attached to every where she appears, so the rubber bands have an equal tug of war".

The resulting picture resembles a map of star constellations but with words replacing celestial bodies.

The patterns at the centre of the concentric ovals reveal significant themes, characters and their positions in the books.

Visual index

Textarc.org is designed as a tool for academics or for people that need to filter a text quickly.

For example lawyers could use the system to search long documents for specific words or characters.

But Mr Paley believes his maps are not a substitute for the original texts.

"This is just a visual index," he said. "I still want you to go back to the real work".