AI Assistance in IDEs and Code Generation Tools – A Comparative Study

Archana K M, Assistant Professor DC School Of Management And Technology, Pullikkanam, Vagamon

AI coding tools within software development have quickly become popular among developers for code work and include Claude, Gemini Copilot, Replit, Amazon Code Whisper, and Tabnine. It also has the possibility of increasing Software Engineering productivity. In this article, we expand on the technologies and benefits of using generative AI in the software industry. Automated code generation tools have made their way into the software development process, giving ways to encode the code by using natural language descriptions or even a fragment of code. Some of these tools include: claude, gemini, copilot, replit, amazon code whisperer & tabnine AI Integration in IDEs It consists of several features as suggestion code based on context, code suggestion/Completion, containing errors, and a set of templates for intelligent code analysis The code generation tool is a form of autonomous generation of specific codes snippets or modules from preprogrammed templates or attributes. This type of research critically evaluates each approach and looks at how they improve developers’ output, help with code generation, and help with code quality assurance. In addition, it reviews the state of flexibility as opposed to automation, the time and usage expenditures of the learning curve of each tool type, and the actual-time help and feedback during the computing phase.

Many institutions are incorporating AI-driven tools in their curriculum, especially in the top BCA colleges in Kerala, ensuring students gain hands-on experience with machine learning and data analytics, Students at DCSMAT College, Vagamon one of the top BCA colleges in Kerala leverage AI-powered coding assistants and data visualization tools to enhance their learning experience and they also proving training sessions for empowering the knowledge about the AI tools.

Input

Dataset 1

get the code for the following:

1. Check the input is palindrome or not
2. Print Fibonacci up to a limit
3. Print the even and odd numbers
4. Print the primes up to a limit
5. Print pyramid with asterisk symbol

Dataset 2

1. Binary Search
2. Matrix Multiplication
3. Insert, Delete, update and Display Linked List
4. Binary tree traversal
5. 3-dimensional matching in graph

Table 1:  Comparison Study

SERIES:  0-Nothing 1-low  2-medium 3-high ( Fig 1: Graphical Analysis )

The table compares four AI models/systems- Claude, Replit, Copilot, and Gemini across various parameters related to code generation, optimization, and execution capabilities. Claude and Gemini are shown to have average code optimization, normal space/time complexity, high response time, and the inability to run code. However, Gemini can verify the code, while Claude cannot. In terms of generating speed, Claude is faster at 1 second compared to Gemini at 3 seconds. For input types, Claude accepts documents in all formats, while Gemini accepts audio and image inputs. Claude is rated as having high efficiency for simple programs, while Gemini is rated as usual for simple programs. Information is provided for replit and copilot for most parameters. The table highlights Claude and Gemini’s relative strengths and weaknesses across various technical aspects, while data for the other two systems is lacking. In essence, the table provides a technical comparison between Claude and Gemini, two AI models/systems, across several key parameters related to code generation and execution capabilities

Conclusion:

This study highlights the strengths and weaknesses of some of the most popular code-generation tools, providing valuable insights for practitioners. By comparing these generators, our results may assist practitioners in selecting the optimal tool for specific tasks, enhancing their decision-making process. These results suggest that GEMINI has the potential to serve as a reliable assistant in programming code generation and software development. Replit’s strength lies in its comprehensive development environment, offering a range of features from code editing to deployment, all within a browser-based interface. Its seamless integration with various languages and frameworks simplifies development workflows, making it an ideal choice for beginners.

References:

[1]Gemini: A Family of Highly Capable Multimodal Models [PDF available],Rohan Anil et al. (You can find the full list of authors in the paper), arXiv (search for the title or arxiv identifier arxiv:2312.11805)

[2]Pethigamage Perera1 , Madhushan Lankathilake2, “Preparing to Revolutionize Education with the Multi-Model GenAI Tool Google Gemini? A Journey towards Effective Policy Making” , Journal of Advances in Education and Philosophy, 2023.

[3] . Hetvi Patel, Kevin Amit Shah and Shouvick Mondal, “Do Large Language Models Generate Similar Codes from Mutated Prompts? A Case Study of Gemini Pro” , ACM ISBN 979-8-4007-0658-5/24/07 , 2024.

[4]Stefan W. Hamerich1, Ricardo de Cordoba ´ 2, Volker Schless1, Luis F. d’Haro2,  “The GEMINI Platform: Semi-Automatic Generation of Dialogue Applications”, ISCA Archive http://www.isca-speech.org/archive, 2004.