Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

From Computational to Conversational Notebooks

Thomas Weber, Sven Mayer

TL;DR

This work investigates how to extend computational notebooks with conversational, LLM-powered capabilities, spanning a spectrum from inline code completion to executable NL-driven interactions. It introduces five concrete interface concepts—inline completion, per-cell chat, side-by-side interfaces, conversational notebooks, and executable conversations—and analyzes their benefits and drawbacks. The authors discuss critical issues such as execution order, context management, explainability, and readability, arguing for hybrid, multi-view notebook designs to balance accessibility with transparency. Overall, the paper provides a roadmap for designing and evaluating LLM-assisted notebooks that better blend natural language interaction with executable code workflows.

Abstract

Today, we see a drastic increase in LLM-based user interfaces to support users in various tasks. Also, in programming, we witness a productivity boost with features like LLM-supported code completion and conversational agents to generate code. In this work, we look at the future of computational notebooks by enriching them with LLM support. We propose a spectrum of support, from simple inline code completion to executable code that was the output of a conversation. We showcase five concrete examples for potential user interface designs and discuss their benefits and drawbacks. With this, we hope to inspire the future development of LLM-supported computational notebooks.

From Computational to Conversational Notebooks

TL;DR

This work investigates how to extend computational notebooks with conversational, LLM-powered capabilities, spanning a spectrum from inline code completion to executable NL-driven interactions. It introduces five concrete interface concepts—inline completion, per-cell chat, side-by-side interfaces, conversational notebooks, and executable conversations—and analyzes their benefits and drawbacks. The authors discuss critical issues such as execution order, context management, explainability, and readability, arguing for hybrid, multi-view notebook designs to balance accessibility with transparency. Overall, the paper provides a roadmap for designing and evaluating LLM-assisted notebooks that better blend natural language interaction with executable code workflows.

Abstract

Today, we see a drastic increase in LLM-based user interfaces to support users in various tasks. Also, in programming, we witness a productivity boost with features like LLM-supported code completion and conversational agents to generate code. In this work, we look at the future of computational notebooks by enriching them with LLM support. We propose a spectrum of support, from simple inline code completion to executable code that was the output of a conversation. We showcase five concrete examples for potential user interface designs and discuss their benefits and drawbacks. With this, we hope to inspire the future development of LLM-supported computational notebooks.
Paper Structure (13 sections, 1 figure)

This paper contains 13 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Extendings Computational Notebooks with LLMs
  3. Executing LLM Responses
  4. In-Cell Coding Assistance
  5. Side-by-Side Interfaces
  6. Conversational Notebooks
  7. Executable Conversations
  8. Advantages and Drawbacks
  9. Inline Auto-Completion:
  10. Integrated Chat:
  11. Challenges of the Notebook Paradigm:
  12. Accessibility vs. Accessing All Information
  13. Future Work

Figures (1)

  • Figure 1: The interfaces with increasing support of LLMs. a) shows the state of the art without LLM support, while b)-f) show the LLM support, which is highlighted in blue for illustration purposes.