Reclaiming the Unexplored in Hybrid Visual Programming
summary
Argues that visual languages currently just mimic textual compilers and demonstrates a prototype hybrid system that stores one semantic model but exposes it as node-and-wire, grid layout as semantic, and a debugging view that shows alternative results (forked evaluation). Prototype animates between views so matching elements stay identifiable, targets concatenative/Forth-like code, and adds a new view in about a day; several views work and one seems likely a dead end.
AI-written summary · about
chapters (14)
- 0:00Hybrid visual programming intro
- 1:04Limitations of visual systems
- 3:35Visual dimensions beyond nodes
- 5:40Underused paradigms visually
- 7:06Benefits of multi-representation
- 10:17Multi-view programming demo
- 14:34Semantic trade-offs in multi-rep
- 16:40Debugging and analysis views
- 18:03Call to explore visual programming
- 19:09Q&A on scaling and recursion
- 21:57Implementation challenges
- 23:23Visualizing imperative and state
- 24:07Restricted-scope views
- 25:09Building a visual programming community
related talks
28mVirtual Domain Specific Languages via Embedded Projectional Editing
31mRethinking Visual Programming
11mSubsequently: Telling stories with pictures makes programs
41mThe Future of Programming
47mReal-time prototyping using visual programming languages
28mAcceptable in the 80s - Revisiting Microworlds
01mAudiovisual Programming in Live Performance
22mA Vision of Miking: Interactive Programmatic Modeling, Language Composition, and Compilation