Foxconn accelerates legacy code understanding by 10x with SAFA.ai

The Story

Maxnerva, a Foxconn subsidiary focused on System Integration, utilizes cutting-edge tools and techniques to help customers integrate and modernize their applications efficiently. To optimize this process, Foxconn partnered with SAFA to automate the assessment, documentation, and analysis of a new customer’s code base. Foxconn’s first step was to quickly understand the functional behavior of the customer's code so that they could identify gaps in functionality or areas for improvement.

The customer provided Foxconn with a Front-End (FE) and Back-End (BE) Codebase.

<aside> 💡 FE Codebase Stats

• Language: React.JS
• 75 Code Files or ~22,500 Lines of Code
• 265 Extracted Functions
• Minimal Comments, Decently Named Variables, No Supporting Documentation

BE Codebase Stats

• Language: C
• 116 Code Files or ~34,800 Lines of Code
• 570 Extracted Functions
• Minimal Comments, Decently Named Variables, No Supporting Documentation </aside>

One of the foundational aspects of any integration and modernization project is the process of creating a comprehensive understanding of the codebase's current functionality and behavior. This allows the team to identify areas for improvements or gaps in functionality. Unfortunately, the provided codebases had no supporting documentation and sparse code comments. The Senior Engineering Manager expected ~30 days to create a holistic understanding of the codebase and its existing functionality, alongside their other responsibilities.

SAFA has automated this process through their Document Generation Platform, and Foxconn leveraged this platform to reduce their time to comprehensive system understanding from 30 days to 3 days, or 10x quicker than traditional tools and processes. By implementing SAFA’s proprietary platform, Foxconn is able to quickly provide value to their customers, and create internal alignment in a shorter period of time.

<aside> 💡 Complete System Understanding from 30 Days to 3 Days via:

• 191 Language Agnostic Code Summaries
• 835 Function/Class Extracted and Summarized
• 2 High-Level Project Overviews for the FE and BE
• 400 Generated Upstream Documents to accurately cluster and describe system-wide functionality
• System-Wide Relationship Mapping and Traceability across generated documentation and code
• Interactive Visualizations and Knowledge Graphs </aside>

<aside> ❓ Check out the documentation SAFA generates on our public codebases page, or try it on your own codebase with SAFA's self-service platform!

</aside>

Challenges

• Sparse to No Existing Documentation or Code Comments
• Needed to identify existing platform functionality quickly
• As digital transformation and system integration initiatives surged, Foxconn sought to expedite service delivery to clients and augment the depth and breadth of analysis provided.

Solution

SAFA’s Document Generation platform allowed Foxconn to extract natural language system documentation using only the code in less than a few hours, reducing codebase onboarding from 30 days to 3 days. This allowed technical and non-technical stakeholders to quickly understand existing system capabilities and create an optimal project path for their customers.