MIP AwardsASE 2022

Software developers spend a significant portion of their resources handling user-submitted bug reports. For software that is widely deployed, the number of bug reports typically outstrips the resources available to triage them. As a result, some reports may be dealt with too slowly or not at all. We present a descriptive model of bug report quality based on a statistical analysis of surface features of over 27,000 publicly available bug reports for the Mozilla Firefox project. The model predicts whether a bug report is triaged within a given amount of time. Our analysis of this model has implications for bug reporting systems and suggests features that should be emphasized when composing bug reports. We evaluate our model empirically based on its hypothetical performance as an automatic filter of incoming bug reports. Our results show that our model performs significantly better than chance in terms of precision and recall. In addition, we show that our modelcan reduce the overall cost of software maintenance in a setting where the average cost of addressing a bug report is more than 2% of the cost of ignoring an important bug report.

Pieter HooimeijerAuthor

Engineering Manager, Facebook Inc.

Westley WeimerAuthor

University of Michigan

United States

no description available

Giriprasad SridharaAuthor

IBM Research Labs

Emily HillAuthor

Drew University

United States

Divya MuppaneniAuthor

Lori PollockAuthor

University of Delaware, USA

United States

K. Vijay-ShankerAuthor

no description available

Julia RubinProgram Co-Chair

University of British Columbia, Canada

Canada

Shahar MaozProgram Co-Chair

Tel Aviv University, Israel

Israel

Marouane KessentiniGeneral Chair

Oakland University, USA

United States

The need for more autonomy in many systems (e.g., automotive, aerospace) leads to increasingly complex systems that are becoming harder to verify and certify. This problem is compounded by the rising use of AI to enable such autonomy in many critical functions. Such situation has a significant impact on software development practice and makes it particularly difficult to assess risks and meet industry standards. This talk will focus on automated techniques, relying on various AI disciplines, to enable the testing and risk analysis of such systems, in ways that are practical and scalable. The talk will report and reflect on various research projects carried out in collaboration with the automotive and satellite industries.

Lionel C. BriandKeynote Speaker

University of Ottawa, Canada / University of Luxembourg, Luxembourg

https://www.lbriand.info

Machines today can write software, compose music, create art, predict events, and listen and learn from humans. Notably, automation also plays an essential role in high performing software development teams by automating tasks and improving developer productivity. But automation can’t (yet) replace human imagination and the intelligence that arises when multiple great minds work together to solve the complex problems that are inherent in software and systems design. In this talk, we will review how automation in modern software development has evolved and the many benefits it has brought. We will then explore how a deeper understanding of the developer experience points to untapped possibilities for innovating automation for software engineering, focusing on how they can:

• support developers to manage the cognitive complexity of today’s systems,
• ease and enhance collaboration by speeding up feedback loops, and
• help developers to get in and stay in a state of flow when developing.

We will conclude by discussing how we can measure the impact of new innovations on the developer experience, and how doing so will drive actionable change and empower developers to do their best work joyfully.

Margaret-Anne StoreyKeynote Speaker

University of Victoria

no description available

Lionel C. BriandChair

University of Ottawa, Canada / University of Luxembourg, Luxembourg

https://www.lbriand.info

Testing has always been about simulation, although often implicit. For example, when a software tester mocks out the behaviour of a code unit, they are effectively stimulating its behaviour. When a test tool creates a test user account in order to log in and run a test it is essentially simulating a potential user behaviour. Traditionally, software engineers, and the research that supports them, has tended to apply greater focus to problems and challenges associated with test mock techniques and applications. However, as software is increasingly deployed to user communities that interact through the software platform, effective techniques for simulating user behaviour are becoming increasingly important. In this talk we outline engineering trade-offs between test (simulation) realism, which we seek to maximise, and various testing costs we seek to minimise. We present an overview of and results from the deployment, at Meta, of two simulation-inspired testing technologies: Sapienz and Web Enabled Simulation, both of which aim to realistically simulate user behaviour, on client side and server side respectively. We show how a simulation-based testing approach opens new perspectives on new and existing automated software engineering challenges and opportunities.

Mark HarmanKeynote Speaker

Meta Platforms, Inc. and UCL

United Kingdom

Inna DvortsovaKeynote Speaker

Facebook

United Kingdom