Hyperglosae: Human annotations for digital humanities

Product vision

For intellectual workers
who interpret documents
Hyperglosae is a digital margin
that allows writing glosses (comments, adaptations or quotes) of these documents for future readers.

Unlike Hypothes.is, Google Docs, Hypotheses (based on WordPress), Zotero, and LiquidText,
our product allows glossing an author-defined "passage" (i.e. a meaningful unit, like a paragraph, a strophe, a verse...).

Unlike Hypotheses,
our product also allows commenting on a reader-defined "fragment".

Unlike Hypothes.is and Google Docs,
fragments may be defined on a picture.

Unlike all of the previously cited software,
fragments may even be defined on a video.

Unlike Hypothes.is, Google Docs, and Zotero,
our product allows glosses to be full-fledged documents and hence to be glossed as well.
It allows navigating from gloss to gloss, or from source to source.

Project background

In the past decades, we designed Cassandre for ethnographic studies and TraduXio for translation studies. It appeared that both pieces of software were kind of implementations of "parallel documents with a visible connection" as envisioned by Theodor Nelson, who coined the term "hypertext". We then figured out that a more general Nelsonian¹ hypertext could be the missing infrastructure for "digital humanists".

Step by step, we will prototype such an infrastructure and test it on practices in social and human sciences.

Requirements

HyperGlosae will be designed with environmental responsibility in mind (for example the number and weight of HTTP requests will be kept low).

Architecture

Bidirectional links cannot be distributed as easily as unidirectional links. If the frontend was the origin of every request (as on the Web), getting all bidirectional links to a given document would require every backend (that may store one) to be queried... Instead, the original hypertext architecture was more like a federation of backends (a bit like Usenet). A modern version of this (see figure below) will be achieved with CouchDB filtered replications.

Deliverables

The folders of the repository correspond to the main deliverables:

• samples of parallel documents, meaningful for stakeholders,
• frontend prototype for reading and writing parallel documents, along with scenarios as user-centered specifications,
• backend prototype for storing parallel documents.

How to test a development version with sample data?

Run the following commands from a terminal (requires Docker and Node.js):

export COUCHDB_USER="TO_BE_CHANGED"
export COUCHDB_PASSWORD="TO_BE_CHANGED"
docker compose --file docker-compose.dev.yml up --detach
cd frontend && npm install
npm start

Open http://localhost:3000 in a browser. To test edit features, log in as user alice with whiterabbit as the password.

How to install a clean stable version?

Run the following commands from a terminal (requires Docker):

export COUCHDB_USER="TO_BE_CHANGED"
export COUCHDB_PASSWORD="TO_BE_CHANGED"
docker compose up --detach

Open http://localhost in a browser.

Footnotes

1. Disclaimer: We are not affiliated with Theodor Nelson. We are just fans ;) ↩