
Potential threats to the validity of LLM-based software engineering research
Large language models have become a popular tool for software engineering research, but they can also influence the outcome of experiments.




Large language models have become a popular tool for software engineering research, but they can also influence the outcome of experiments.

Training and using large language models to develop software is bad for the planet – but it doesn’t have to be that way.

Dark and high-contrast modes help make apps more accessible, but not all support them well enough.

The five most popular JS frameworks – Angular, React, Vue, Svelte and Blazor – use different rendering strategies, and it shows.

A group of researchers studied what happens when you explicitly remind large language models to respect copyright notices.

Who could have thought that DOM-based browser extension fingerprinting can be easily mitigated by splitting reality?

Installing browser extensions can be bad for your privacy, even if they were made in good faith.

This paper provides an analysis of browser extensions in the Chrome Web Store based on historical data and their source code.

Everyone knows that the cost of fixing a bug grows exponentially as it progresses through development phases. But is that actually true?

Applying these design patterns in your code may (or may not) make your web applications more energy-efficient.

Evolving microservice APIs requires a lot of work, both from the providing and the consuming side.

Alternative media platforms and their communities use visual disinformation in a variety of ways.

Most initiatives that use artificial intelligence to combat disinformation are clustered around two continents.

Google’s API Governance process makes sure that the design and development of its APIs meets the company’s standards.

Large language models (LLMs) may not be replacing developers anytime soon, but can still be of use in the design process.

This study looks at how humour can fuel developer engagement in open-source and professional contexts.

Large language models can be used to deduce to what extent your system’s implementation still adheres to a reference architecture.

This study pits 14 open-source main content extractors against each other and arrives at a somewhat surprising conclusion.

End-to-end tests can help you discover problems in web applications, but sadly are not free of problems themselves.

Supersharers managed to reach 5.2% of registered voters. In contrast, Russia’s 2016 campaign only reached 3.4% of voters.

Although predatory journals may sound scary, they are actually quite friendly (and after your money).

Public service broadcasting has served us well in the 20th century. How can it keep serving us well (well) into the 21st century?

Sharing is caring, but if you really care about sharing research you should probably do more than just share links to papers.

New live broadcasting technologies have transformed how news organisations work, in both good and bad ways.