Whither Problem-Solving Environments?
summary
Argues that Jupyter-style notebooks and IPython give easy access to languages, libraries and visualization but leave gaps: poor problem-domain interaction, fragile reproducibility, no experiment management, weak component integration and awkward deployment. Literature on problem-solving environments rises through the 1990s then publication counts drop after 2005; an analysis of over a million notebooks shows many lack narrative. Describes prototype work using conceptual modeling and diagrams-as-code, and recent small experiments with LLMs; results are promising but incomplete.
AI-written summary · about
chapters (10)
- 0:00Problem-solving environments intro
- 1:04What happened to problem-solving environments
- 3:10Engineering knowledge from PSE
- 7:26PSE key features
- 18:00PSE research analysis
- 20:49PSE research rise and fall
- 24:02Notebooks as PSEs
- 30:04Notebooks vs PSE criteria
- 34:40PSE future directions
- 38:46Q&A
related talks
33mSignificant Roadblocks to Usefulness for Jupyter Notebooks and a Recipe to Fix them
25mOvercoming 5 Hurdles to Using Jupyter Notebooks for Data Science, by the JetBrai...
17mNotes on Notebooks: Is Jupyter the Bringer of Jollity?
45mSolving problems by sharing them... with Python!
25mThe First Notebook War
41mBetter Code for Data Science
21mThe Story in the Notebook: Exploratory Data Science using a Literate Programming Tool
01h 00mInteractive Computing with Jupyter: Past, Present, and Future