Algorithms, avocados, and adventures

The step abroad was a deliberate adventure for Professor Sebastian Erdweg. When planning his research sabbatical, the computer scientist of Johannes Gutenberg University Mainz (JGU) envisioned not just an academic collaboration, but also family life in an unfamiliar country. As a visiting professor, he eventually went to the Universidad de Chile – bringing home not only formative experiences but also a major third-party funded project.

"We have two young children, my wife was on parental leave at the time, and we wanted to combine the opportunity of a research sabbatical with an extended family trip," recounts Professor Sebastian Erdweg, head of the Programming Languages research group at Mainz University. For a ten-week guest professorship, he and his family visited the Universidad de Chile, one of South America's leading research institutions in this field. "We traveled from September to November 2020, which is springtime in Chile," Erdweg explains. "The trip was meant to provide shared experiences for my family – and for me the opportunity to better connect with my colleagues and their work."

Erdweg's team at JGU develops new programming languages and tools designed to help developers create reliable as well as efficient software. The group specializes in three key areas: domain-specific languages (DSLs), tailored for specific application domains; static program analysis, which identifies errors before the program is run; and incremental computing, which re-executes only the parts of a computation that have changed. "This saves time and resources, especially for large datasets or complex applications," says Erdweg, who recently received an ERC Consolidator Grant for this research.

Glaciers and guanacos

Erdweg's decision to spend his sabbatical in Chile was not based on the typical content-driven expectations. "Unlike most projects, this one didn't have a specific academic imperative," he explains. "Together with my Chilean colleagues, computer scientists Professor Éric Tanter and Professor Matías Toro Ipinza, I simply wanted to explore our joint scientific potential – without the pressure to deliver immediate results." While in Santiago de Chile, Erdweg presented two guest lectures and several research talks. "It quickly became clear that our interests complement each other in many ways."

The Mainz team's expertise in predicting program behavior through abstract interpretation aligned well with the Chilean researchers' work in gradual typing, which checks for errors both before and during program execution. Combining these areas of focus, they concluded, could yield a novel approach to interconnected and theoretically sound hybrid analyses, enabling even more precise error detection and program optimization.

In this process, the in-person exchange proved invaluable. "In computer science, as in most disciplines today, much can be done remotely," Erdweg emphasizes. "But only personal meetings allow you to understand each other's working styles and build trustful relationships. In academia, unlike in companies with a unified work culture, teaching, research, and infrastructure often differ between universities." Trust fosters creativity and the necessary openness to discussing unpolished ideas. "In our field, every single publication is essentially an 'invention' to be safeguarded. At the same time, innovation needs creative exchange. Trust is crucial in all this – and something that cannot be built easily during online meetings," emphasizes Erdweg.

The stay did not only bring academic achievements, but also many personal experiences. "In Chile, we cooked, shopped, and attended street festivals – even in non-touristy areas," Erdweg says. "It gave us deeper insights into the country and allowed us to immerse ourselves in the culture." The family, with children aged 10 months and 2 years at the time, explored the country during their free time. "We went on excursions almost every weekend," he recalls. "A highlight for the kids was a hike to a glacier in the Andes, another was a trip to Patagonia." The family was captivated by guanacos, wild relatives of llamas roaming freely through the rugged landscape. "And avocados, which are abundant in Chile, quickly became a family favorite." It was a truly formative experience. "My son, who turned three in Chile, learned a lot – like how to communicate with local kids without sharing a common language."

Tin huts and banking district

The trip also opened the family's eyes to social disparities in Chile. "There is a gap between rich and poor that we don't experience to this extent in Germany: on one side, corrugated tin shacks, on the other, a glittering financial district." These differences are also reflected in university life, where PhD students often rely on scholarships and, due to high rents in the capital, frequently live with their parents. "These disparities also explain why security measures at the University of Chile are much stricter than at many European universities," Erdweg notes. "A security team closely monitors campus access, and special ID cards are required to enter university buildings."

While this was unfamiliar to Erdweg, it also provided a sense of safety – especially after an incident on the very first day. "While exploring downtown with my wife and kids, my phone was stolen from my pocket." However, the family's sense of security quickly returned, "thanks to our truly wonderful experiences and my colleagues' hospitality." Éric Tanter frequently invited the family to join barbecues in his garden, for example.

This personal trust ultimately paved the way for a long-term collaborative research project. "Just two weeks after my return to Germany, our teams saw a Joint Call for Proposals from the German Research Foundation (DFG) and the Chilean National Research and Development Agency (ANID) – and we decided immediately to apply," Erdweg recalls. "Thanks to the intensive weeks we had spent together, we were able to demonstrate an established collaboration – a rarity in such applications."

Embracing the unknown

The joint project now underway aims to develop new methods for reliably detecting programming errors. "Using the programming language WebAssembly as an example, we are working to improve program analysis," explains Erdweg. Together, he, Tanter, and Toro are combining their expertise in abstract interpretation and gradual typing to create a novel analysis method. "The goal is to identify security vulnerabilities at an early stage and make program execution safer and more efficient."

With a total budget of EUR 450,000, of which EUR 350,000 is allocated to JGU, this is the first joint computer science project between the two universities and a direct result of Erdweg's research sabbatical. "The freedom to do research without a specific goal and the personal encounters on-site made this possible," he emphasizes.

The project has already seen some success. In June 2024, Chilean colleagues spent two weeks in Mainz, advancing the work and deepening the collaboration. "Much can grow from this – joint publications, PhD exchanges, and surely more projects," Erdweg says. He hopes his experience will inspire other researchers to undertake similar international journeys. His advice: "Be open to the unknown and life-enriching new impressions."

Related links:

• Professor Dr. Sebastian Erdweg
• Programming Languages research group at the JGU Institute of Computer Science
• Institute of Computer Science at JGU

Read more:

• press release "Computer scientists of Mainz University and the University of Chile cooperate to establish a reliable detection of programming errors"
  (11 June 2024)
• press release "Reducing the energy consumption of software: Sebastian Erdweg receives an ERC Consolidator Grant"
  (5 Dec. 2023)