FFL: A Language and Live Runtime for Styling and Labeling Typeset Math Formulas

Document Title

FFL: A Language and Live Runtime for Styling and Labeling Typeset Math Formulas

Document Information

• Domain: Tools for annotating and presenting mathematical formulas
• Keywords: Mathematical formulas, interactive typesetting, real-time code feedback, annotation, color enhancement

Research Background and Problem

• Research Problem or Challenge: The complexity of mathematical symbols and formulas often hinders their comprehension by non-expert readers. In fields such as machine learning or data science, key formula components in literature frequently lack intuitive visual annotations, making it difficult for readers to quickly grasp their meaning. Existing formula typesetting tools (e.g., LaTeX) have complex marking systems and do not support flexible cross-document style modifications.
• Importance: Improving formula annotation is crucial for education, research, and readers' knowledge acquisition. Intuitive and convenient tools for enhancing mathematical formulas can help a broader audience understand complex formulas.
• Motivation and Related Work: Existing research indicates that color coding and annotations for formulas can reduce cognitive load. However, current tool designs still have limitations in usability, real-time feedback, and general applicability. Recent web tools like KaTeX provide only preliminary support for style control and struggle to effectively integrate large-scale interactive style enhancement needs.

Solution

• Method or Solution: Developed a markup language named FFL (Formula Formatting Language) and its accompanying live runtime environment. FFL mimics the structure of CSS, separating style definitions from formula structures to add styles and annotations to formulas in a more readable and modular way.
• Innovations:
  1. Introduced a selector-based style rule system that can precisely target specific expressions within formulas.
  2. Supports dynamic label generation and style preview, significantly enhancing real-time feedback.
  3. Developed a markup parsing tool that converts LaTeX formulas into stylizable HTML, allowing users to flexibly customize formula enhancement styles.
• Implementation Steps and Key Technologies:
  1. Provided an FFL parser that allows users to define formula styles (e.g., color, borders, and text weight) using selectors and style rules.
  2. Adapted user modifications in real-time, rendering styles as HTML + CSS through integration with the KaTeX engine.
  3. Developed an automatic label layout algorithm to optimize label positioning and prevent overlap, reducing the need for manual adjustments.

Research Outcomes

• Specific Results:
  • FFL outperformed LaTeX in readability and style editing efficiency. In experiments, participants completed complex formula editing tasks faster with FFL than with LaTeX, while error rates decreased.
  • Achieved cross-document, consistent style update functionality, effectively supporting more complex mathematical text creation.
  • FFL offered convenient style reuse features, with user feedback indicating more intuitive operation.
• Comparison with Existing Solutions:
  • Formula annotations using FFL are independent of the formula body, avoiding the issue in LaTeX markup language where adding or removing styles disrupts formulas.
  • FFL provides instant style previews, significantly reducing the cycle for style debugging and validation.
• Experimental or Evaluation Results:
  • In author efficiency experiments, participants edited complex formula styles approximately 30%-40% faster using FFL compared to LaTeX.
  • For simple tasks, there was no significant time difference between FFL and LaTeX, but FFL scored higher in user experience evaluations.
  • User surveys showed that 82% of respondents found FFL's multi-style synchronization and real-time feedback features particularly useful.
• Limitations and Future Directions:
  1. Current implementation has limited compatibility with LaTeX formula markup matching; future improvements could leverage AST (Abstract Syntax Tree) integration.
  2. Error recovery and suggestion functionalities are still at a basic level and require more user input and automation support.
  3. Future extensions could include dynamic formula interactions (e.g., value adjustments and real-time propagation calculations) and animated formula presentations.
  4. Optimize accessibility for blind or visually impaired users (e.g., enabling screen readers to read formulas and their annotations).