Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Help the Whitney Museum recover a web-based artwork from the 1990s

In 1995, Douglas Davis's The World's First Collaborative Sentence (1994) was exhibited at the first Gwangju Biennale, where visitors were able to contribute to the piece using machines provided by Samsung. Most contributed text in their native languages, and as a result, approximately 40% of the Sentence is in Korean. Unfortunately, before the work was donated to the Whitney Museum, the files sustained some form of encoding damage. All of the Korean text input by visitors at the 1995 Gwangju Biennale is now illegible.

For example, what once would have been seen as: 광주비엔낱레여 (which Google translates to "Gwangju Biennale open"), is now rendered as ±Ûç…¼–À£Òøá. Some modern browsers attempt and fail to automatically detect the encoding. This is not a simple matter of the text being in an obsolete character encoding, but rather, it would appear that the text was inadvertently converted between character encodings. This could have been through some act of copy and paste, but the details of how or when this damage occurred are unknown.

The challenge

The Whitney Museum has conducted initial research and experiments in encoding mapping and conversions, and has assembled all available evidence here in this repository. We're not sure if complete recovery is possible, and in any case it is beyond our area of expertise. Here's where you come in.

The challenge: Repair the damaged character encoding to its original state. It is required that your results be reproducible. We ask that you produce technical documentation of your process and solution, explaining the "how" and "why" of what you did.

Your tools

  1. The Evidence directory contains a PDF containing the ONLY remaining evidence of the correct rendering of the Korean text. We have also included a snippet of the source files to which this excerpt corresponds.

  2. The Research directory contains correspondence between Ken Lunde (author of CJKV Computing [from O'Reily]) and Ben Fino-Radin, in which they discuss the root of the problem, and potential solutions.

  3. The Source Files directory contains all pages of "The Sentence." Most of the Korean text deriving from the Gwangju Biennale can be found in files sentence5.html – through – sentence13.html

To respond to the challenge

  1. Fork the repository

  2. Do your magic

  3. Email a link to your solution and documentation to:

It's worth it

If you are able to repair this encoding damage, this will be the first time these characters have been properly seen by anyone in at least 15 years.