Research

January 6, 2026
in Artifact Evaluation, DSLs, Helical, Research
6 min read

Challenges with Locality and State in ExPL Annotations, Part I

ExPL, the experiment specification language of Helical, relies on user-provided annotations to infer an implicit post-interventional causal structure and query set. Annotations are attached to ExPL expressions and provide a link to the program variables in HyPL. Because ExPL encodes stateful executions, we need to be able to reason about the scope over which stateful operations apply. In this two-part blog post we'll talk about some challenges associated with using these annotations.

November 7, 2025
in Artifact Evaluation, Reproducibility, Research
9 min read

POPL 2015 Artifact Evaluation Revisted

Just over a decade again POPL ran its first artifact evaluation and I was on it! I'm listed in the proceedings as "Emma F. Tosch" as an inside joke with Arjun, who ran the committee. I have what only I think is a Very Interesting Tale of reviewing, but that's not what this post is about. Instead, this post is about tracking down those old artifacts with just a pinch of oral history.

February 9, 2023
in Research, Reproducibility
7 min read

Reproduction vs. Replication

In 2020 or so, ACM swapped its definitions of reproducibility and replicability to be more in line with the broader social sciences community. I see the reasoning for the swap, but I don't think that the new definition is quite right either. The crux of the issue is that definitions both communities are using are a shorthand that doesn't map appropriately across communities.

August 18, 2022
in Research
5 min read

Non-cooperation in experiments

Last week, Kaleigh Clary presented her work, Stick it to the Man: Correcting for Non-Cooperative Behavior of Subjects in Experiments on Social Networks at USENIX Security. One of the things I really like about this work is the focus on analysis and implications of non-cooporative actors in network experiments from a PL/systems perspective.

Kaleigh Clary presenting her work. Slide depicts the scenario as a causal model, where non-cooperative behavior is defined as the case where participants behave differently when they believe they are being surveilled.

July 20, 2022
in Research
5 min read

Answering the important questions

Here in the MaPLE lab, we treat the techniques students cultivate in their programming systems classes as general methods for answering a broad range of research questions that arise in and around software when it is used in nontraditional ways. As a result, students without formal training in the techniques we use often ask: "What is PL?," whereas those with training often ask: "How is what you do PL?"

You must not know about me