On the net, the starting point in reading a text is almost arbitrary and driven by natural language processing and information retrieval tools. On the other hand branching out from any point of text is inevitable. TEWXT is an experimental reading/writing tool that realizes this experience within a web page.

TEWXT creates a hypertext structure from a corpus of text not by using hype-links, but by allowing users to branch-search-read at any point.

In TEWXT users can use any of the search boxes on the screen to search for a word within a corpus of texts, choose one of the results and start reading the text. At any point users can open a new box in any of the adjacent positions, perform another search and follow a new path in reading. Also it is possible to open new boxes for marginal notes.

This structure of text boxes can grow infinitely.

It is possible to record and replay the whole experience of reading (including notes) using TEWXT.

Using each box (or combination of adjacent boxes), users can apply natural language processing algorithms to the content of the boxes and transform and create new text:

  • Wikifying: highlighting and linking words within the text to Wikipedia articles by a natural language processing algorithm.
  • WIkiTransformation: Transforming text by substituting words within text using Wikipedia and the semantic relevance of the linked articles as thesaurus.
  • Markov Text Generation: Generating new text by applying Markov Chain Text Generation Algorithm to text inside a box.
  • Cellular Automata Generation: Using adjacent text boxes as cells of a cellular automata to generate new text.

Moreover users are able to combine retrieved text, notes and transformed text to create a new text which implicitly resonances the reading trails being used as a high-tech tool to create something like Walter Benjamin’s ‘Arcade’ or just go through ‘The garden of forking paths’

← Back