hyperpad
A peer-to-peer collaborative text editor for humans and communities.
What is it?
Hyperpad is a free, open source, peer-to-peer text editor for people and their communities. Authors control who gets access, and data is hosted by the peers who are interested in it.
Current Status
I've paused work on the web client in favour of building out an Electron-based native client, hyperpad-desktop. This lets me dodge some difficult problems around web tech (indexed-db performance; webrtc) and focus on making the core hyper-string primitive robust and durable.
Why another collaborative editor?
Some of the most popular collaborative document editors today include Google Docs and Etherpad.
Google Docs gets the fundamental piece right: real-time text editing. However, all of your data is stored by and readable by Google, Inc. It is closed source proprietary software.
Etherpad takes this a step further in multiple directions: it is open source, and can be deployed by anyone on any server. This lets any individual or group run etherpad and keep ownership and privacy to their data.
Etherpad is most of the way there, but Hyperpad goes the rest of the way in two crucial aspects:
1. No servers required
In peer-to-peer networks, all users are equal.
Nobody needs the monetary resources and technical know-how to run a server.
Unlike centralized services, you own each pad you create. Turn on encryption, and your data becomes unreadable to anyone but those you grant access to. There are no service providers to go out of business and lose your data.
Everything is client-side HTML and Javascript: you can just save the Hyperpad website and run it locally on your computer, and it will function just fine!
2. Works great offline
Not everybody in the world is online. Among those who are, many do not have consistent, broadband connections. People-respecting software must work excellently offline; no exceptions.
Forgetting this is the The Silicon Valley Privilege (TODO: link to article).
Hyperpad uses an eventually consistent data structure called hyperlog, which operates happily offline and will sync with other users whenever a network connection is available.
How does it work?
Hyperpad relies on the browser itself for storing documents, and powerful peer-to-peer primitives like WebRTC and hyperlog to organize and transfer documents to those with access.
The act of having a document open in your browser immediately lets a user act as a host for that document's data, sharing it in real-time with others with others. In the case that a user is offline, they can still freely make edits locally, which will propagate to others storing the document when they re-establish a network connection.
Hyperpad is built in a modular fashion atop a set of do-one-thing-well modules:
- hyper-textarea: back a textarea with a hyper-string for conflict-free p2p replication
- hyper-string: a conflict-free p2p string data structure
- textarea-op-stream: readable stream of a textarea's inserts and deletes
- lots of great modules from mafintosh: hyperlog, signalhub, and others!
Coming Soon(tm)
- hyperpad-desktop
Faster operations (batching in thehyper-stringlayer)- Encryption (with separate read/write privileges)
- Secure app delivery (maybe hyperboot)