The Reserve Reviewer Policy
The Reserve Reviewer Policy is an experimental policy to help conferences scale to increased workloads. We outline its design and first implementation.
AI Models Need a Virtual Machine
Neural networks are more useful when placed in a suitable, specialized environment.
The Academic Pipeline Stall: Why Industry Must Stand for Academia
This post was cross-published from the SIGARCH blog. The Research Pipeline is Stalling The U.S. National Science Foundation (NSF) froze all outgoing funding, including new awards and scheduled payments on active grants. Over 1,000 NSF research...
How to Give a Good Talk
Testing AI Software Isn’t Like Testing Plain Old Software
AI software demands new approaches to testing that go far beyond existing software testing methodologies. And with the rapid evolution of AI model capabilities, the need for the software engineering community to innovate new approaches to AI software testing is urgent.
Parametric Subtyping for Structural Parametric Polymorphism
Recursive types, generics (sometimes called parametric polymorphism), and subtyping are all essential features for modern programming languages across numerous paradigms. However, structural subtyping is undecidable in the presence of recursive types and generics. In our POPL 2024 paper and its accompanying implementation, we propose a reconstruction of the interaction between recursive types, generics, and structural subtyping from first principles. We present a notion of parametricity for type constructors that forms the basis of a suitable, decidable fragment of structural subtyping, which we call parametric subtyping.
The Missing Mentoring Pillar
“Should PLDI return to in-person Program Committee meetings?” – Survey results
Evaluating Human Factors Beyond Lines of Code
Software systems researchers want to make human-centered claims, but don’t have the proper tools to do so. That’s how we ended up with the ubiquitous lines-of-code comparison found in evaluation sections everywhere. In this post, I explore: what can we reasonably expect researchers to do that’s higher quality than LOC but less effort than a full-blown user study?
