Sonoma State University
Department of Computer Science
CS-370: Software Design and Development
Exercise 24: Artificial Intellgience Driven Software Development

Objective:

To assess the effectiveness of Artificial-Intellgience-driven software development tools.

Exercise:

Now that you have experience developing a web-based auction website and know the security vulnerabilities and all sorts of gotchas that plague web development, I'd like you to "vibe code" with ChatGPT.

Specifically, I want you to prompt ChatGPT to create an online auction website like just like the project you developed in our CS-370 class. This process may require multiple prompts and iterations to finally get a software project suitable for evaluation. I suggest prompting ChatGPT for no more than one hour so that you will have adequate time to evaluate the results.

Next, I want you to evaluate the output source code generated by ChatGPT by answering the following questions:

  1. What programming language(s) did ChatGPT use?
  2. Is the source code easy to read and understand?
  3. Does the source code appear complete and fully-functional?
  4. Are there any security vulnerabilities in the source code?
  5. Does the source code require any external dependent libraries?
  6. Does the source code require use of a database management system? Which one?
  7. If a database management system is implemented and it uses Structured Query Language (SQL), does the structure of the tables appear to be in first-normal form, second-normal form, or third-normal form?
  8. What is your overall assessment? Is AI-driven software development the future of software development? Justify your answer.

Uploading your solution:

  • This is a group exercise. Please list all team members who worked on this exercise. Only students whose names appear on the exercise will receive credit for this exercise.
  • Please upload your answers to the above questions in PDF (Acrobat), Word (doc/docx) or a text file.
  • Note: You do not need to upload all the source code generated by ChatGPT.