Sample paper
University: Federation University Australia
Word Count: approximately 3,500 words
Language Name
The proposed language is named Cogent.
Introduction and Language Purpose
Cogent is motivated by AI's status as a relatively new and dynamic field focused on building machines with human-like abilities -- vision, speech, cognition, and language translation -- which the essay argues imposes distinct requirements on the languages used to build such systems (Russell & Norvig, 2021). Cogent is proposed as a response to these requirements, designed specifically to support the construction of AI systems rather than adapted from general-purpose origins.
Interpretation/Compilation Methods, Memory Management, and Scoping
Following the structure used across this course's language-design assignments, Cogent specifies a compilation strategy referencing LLVM-based native code generation (Lattner, 2008), consistent with a hybrid interpretation/compilation approach balancing interactive development against production performance. Its memory management design draws on established garbage collection and region-based memory management research -- generational scavenging (Ungar, 1984) and static region-based management (Tofte & Talpin, 1997) -- reflecting a deliberate grounding in established memory-management literature. Scoping and resource-safety concerns are informed by formal language-theoretic work on linear types (Wadler, 1990) and algebraic effect handling (Plotkin & Pretnar, 2013), suggesting a restricted-aliasing and exception-handling design with a stronger theoretical foundation than a purely pragmatic, engineering-first approach would provide.
Language Features and Rationale
Cogent's specification addresses the nine major language-design features common to this course's assignments -- simplicity, orthogonality, data types, syntax design, abstraction support, expressivity, type checking, exception handling, and restricted aliasing -- each evaluated against the demands of AI system development. The essay's engagement with foundational language-design literature, including Wirth's (1974) work on the design of programming languages and Milner's (1978) theory of type polymorphism, situates these feature choices within the established academic tradition of programming language theory rather than treating them purely as engineering trade-offs specific to AI tooling.
Readability, Writability, and Reliability
The essay closes by assessing how Cogent's design choices -- its LLVM-based compilation strategy, its academically grounded memory management approach, and its type-theoretic foundations for aliasing and exception handling -- jointly support readability, writability, and reliability for AI codebases, following the same evaluative framework applied across this course's language-design proposals while distinguishing Cogent through its closer engagement with formal programming-language-theory literature.
Conclusion
Cogent is presented as a programming language for AI systems development grounded more explicitly than typical engineering proposals in established programming-language-theory research, from generational garbage collection and region-based memory management to linear types and algebraic effect handling. This theoretical grounding is argued to give Cogent's design choices a stronger formal foundation for supporting the reliable, maintainable construction of AI systems with human-like capabilities in vision, speech, cognition, and language translation.
References
Milner, R. (1978). A theory of type polymorphism in programming. Journal of Computer and System Sciences, 17(3), 348-375. Plotkin, G. D., & Pretnar, M. (2013). Handling algebraic effects. Logical Methods in Computer Science, 9. Russell, S. J., & Norvig, P. (2016). Artificial intelligence: a modern approach. Pearson. Sebesta, R. W. (2004). Concepts of programming languages. Pearson Education India. Lattner, C. (2008, May). LLVM and Clang: Next generation compiler technology. The BSD Conference (Vol. 5, pp. 1-20). Tofte, M., & Talpin, J. P. (1997). Region-based memory management. Information and Computation, 132(2), 109-176. Ungar, D. (1984). Generation scavenging: A non-disruptive high performance storage reclamation algorithm. ACM SIGPLAN Notices, 19(5), 157-167. Wadler, P. (1990, April). Linear types can change the world! Programming Concepts and Methods, 3(4). Wirth, N. (1974, August). On the design of programming languages. IFIP Congress (Vol. 74, pp. 386-393).