ΤTTandem Texts

How Tandem Texts works

Where the originals come from

The Greek and Latin texts come from two open scholarly projects, the Perseus Digital Library and Open Greek and Latin. Each work names the printed edition behind its text. We never change the original; where the digital text contains a slip, a textual note explains it and we report it to the source project.

Who translates

Every translation starts as a draft written by an AI model that follows a published style guide and glossary, so that a term like ousia is translated the same way across an author’s works. Each passage records which model drafted it. Drafts can contain mistakes: treat them as a first reading and check important passages against a critical edition.

How readers improve it

Anyone with an account can suggest a better wording, with a sentence explaining why. Accepted suggestions earn five reputation points; rejected ones cost one. At 15 points you edit directly, and at 50 points you review other people’s suggestions and can restore earlier versions. Moderators handle disputes.

Every passage has its own discussion, where readers argue about wording and notes before or after an edit.

Every word has an author

The site records who wrote each word of each translation, across every version. Switch on “Who wrote what” in the reader to see it. Words keep their author when others edit around them, and deleted words that are restored go back to whoever first wrote them. Credits for each work are calculated from these shares.

Difficulty and impact

Difficulty is measured from the original: how long the sentences are and how much of the vocabulary lies outside the author’s most common words, scaled from 1 (introductory) to 5 (expert). Impact rates how central a work is to its field and to later tradition, from 1 (specialist) to 5 (canonical); for now it is an editorial judgement.

Languages

The interface is available in English, German, French, Italian and Spanish. Translations of the texts are added language by language; where your language is not available yet, the English translation is shown and marked as such.

Privacy

We count which texts are read and for how long, so we know where to put effort. When you leave a page it sends one small report: the page, how long it was open, which passages were on screen and for how long, the reading mode, the interface language, your country, whether you used a phone and the site that sent you. There are no cookies, no IP addresses are stored, and nothing links one visit to another.

If you create an account, your browser keeps a private key so you stay signed in. We store your chosen name, your contributions and your reputation; contributions are public.

Licences

The texts and everything built on them, including translations, notes and discussions, are shared under Creative Commons Attribution-ShareAlike 4.0. You may copy and adapt them if you credit the contributors and share your version under the same licence. By contributing you agree to release your contribution under this licence. The software is open source under the GNU Affero General Public License 3.0.