Collaborating With The Invisible Talk at replit
Talk given at FISH at Replit.
GOAL:
- Help you see what's in plain sight - give you a framing for thinking about developing collab experiences
- Prompt you to generate ideas for the repl experience
Who am I?
- Jess Martin
- Research into better computers
- Real-time collaboration
What's the problem?
- Invisible Collaborators Problem
- Look at a real interface: Trello
- This sucks!
Why does this even matter?
- Collab was part of the vision from the beginning (1968)
- Mother of all demos, etc
While Engelbart’s NLS platform is now mostly celebrated for introducing the first computer mouse and elements of hypertext and graphical interfaces, the system also offered an early glimpse of what would become known as collaborative software. (Engelbart is sometimes referred to as the “father of groupware.”) NLS supported teleconferencing, shared files, and online communities. In the decade after Engelbart’s famous demo, the mass market success of personal computers convinced Engelbart that something crucial had been lost. “The personal computer revolution,” he wrote in the 1980s, “turned its back on those tools that led to the empowering of… distributed work groups collaborating simultaneously and over time on common knowledge work.” It wasn’t until the rise of cloud computing and services like Slack and Google Docs that Engelbart’s original vision of collaborative software truly came of age.
- The better computer will be collab
- Best way to think of AI is as another collaborator
- Need all the same affordances
How did we get here?
- Collaboration is a Spectrum
- At least 4 modes
- The technology was hard, so we went after single-player
- Eventually, the tech got better
- Croquet, Liveblocks, roll your own
- Examples
- CRDTs, replicache
- Croquet, Liveblocks, roll your own
- UX hasn't caught up!
Where do we go from here?
- "Consider a room!"
- First story suggests a good direction to pursue: start by asking "how do we do this in the real world with physical objects?"
- Yiliu working in a design studio leading to working on AR
- Whiteboarding
- First story suggests a good direction to pursue: start by asking "how do we do this in the real world with physical objects?"
- Communicate intent: what does this person intend to do?
-
Embodiment matters - we have to externalize
-
Replit can change the world
- Fewer sunk costs in re-inventing the IDE
Why does this matter?
-
real-time collab is happening all the time - impoverished
-
Invisible Man Board Game example
-
Let's look at an interface: Trello
-
Collaboration Spectrum:
-
Let's look at replit
Story: working with a real-time framework. The tech advanced, but the UX hasn't caught up.
- Integrally related to embodiment. (Mediums)
- Have to have an external artifact to "think with"
- current IDEs were built and optimized for fully async case
- Won't get to good "fully sync" experiences by starting with async
Not a surprise that spatial canvases took off when things went remote: easy way to quickly externalize unstructured ideas so they can be discussed
History of computing:
- Collaboration has been part of the vision from at least the mother of all demos
- local-first software paper - describe the research goals
Examples of environments with dynamic collab: - whiteboards
- Map room
Computing as a room