Editorial: ICSE and the Incredible Contradictions of Software Engineering

Once upon a time, we submitted our papers to the International Conference on Software Engineering (ICSE) at the end of August and presented them in May, and the community was debating about the long 7-month delay between the submission and presentation of new ideas in a fast-changing discipline such as software engineering.

Today, we submit our papers to ICSE in March and present the new ideas in April... oops... in April 13 months later! In 2024, we meet at ICSE in April to discuss papers that we submitted in March 2023, letting the new ideas of the papers already submitted in March 2024 age for 11 more months up to the presentations at ICSE in April 2025. At the same time, we discuss the incredible acceleration of ideas, technology, and society.

Software is the core component of the current technological revolution: AI is software, machine learning is software, virtual reality is software, the internet of things is software, the whole environment is, by and large, software. Software engineering is a key discipline in this fast-evolving time, and it has the main responsibility of shaping the leading edge of the current technology revolution. There is an obvious and alarming gap between the fast-evolving time of the technological revolution and the fast-growing delay of sharing new ideas and results.

The main software engineering conferences have done an amazing job to improve the quality and consistency of the review process and offer excellent programs while handling a number of submissions that exploded in the past decade. The two submission cycles both distribute the huge review workload over a long period and improve the reviewers–authors interactions, thus enhancing fairness and consistency. Unfortunately, the timeframe between the first submission and the presentation of a paper dilates beyond the interval frame between two consecutive conference editions, leading to the paradox of presenting ideas and results 13 months after the initial submission, while technology changes on a monthly basis.

A decade ago, the main software engineering journals were suffering a crisis of submissions. Back in 2014, ACM Transactions on Software Engineering and Methodology (TOSEM) received as few as 146 original submissions, one-third of the 496 submissions to the ICSE 2014 technical track. None of the software engineering journals was in the first Scientific Journal Rankings (SJR) quartile in 2017. The main software engineering conferences and journals worked together to counteract the decline of journals, and they introduced the Journal First practice that reverted the trend and brought healthy growth back to software engineering journals that has been beneficial for both journals and conferences. In 2023, ACM TOSEM received 495 original submissions, about 65% of the 780 submissions to the ICSE 2014 technical track, and the trend in 2024 is still positive. The three major software engineering journals are now back in the first SJR quartile for the “software” category, despite maintaining stable review turnaround times and acceptance rates over the years.

ACM TOSEM has addressed the needs of timely publication of new ideas with the fast impact track that aims to cap the review turnaround time and accept papers within 6 months from the initial submission, including a major revision. TOSEM has multiplied the issues per year from four issues until 2022 to eight issues in 2024, to reduce the waiting time of accepted papers to appear in an issue. The new “Frontiers of Software Engineering” track hosts new ideas, research directions, and groundbreaking results that may not be supported by solid experimental results yet. We aim to review Frontiers of Software Engineering papers within 60 days and publish them soon afterward, to give fast visibility to new ideas in a quickly evolving epoch.

More and more often, authors publish preprints in open access repositories and present new ideas and results at open webinars. Open access repositories and webinars are great tools; however, they cannot substitute peer-reviewed conferences and journals that offer invaluable feedback to improve the presentation of new ideas and results and skim the unmanageable amount of new publications that flood the Web. Both journals and conferences can play an important role but cannot fully solve the dilemma of thoroughly reviewing while presenting new ideas and results in a timely manner with isolated initiatives. We need a joint effort to optimize and coordinate the important review effort to present all valuable ideas and results in a timely manner in a fast and healthy growing community, which daily produces an enormous amount of valuable ideas and results. We as a community shall rethink the way we share ideas and go back to the straight path before it is too late.

The well-established Journal First initiative offers the opportunity of multiple review cycles to refine papers before presentations at main events without strict deadlines. The newborn Software Engineering Sustainable Community Review Effort (SCRE) initiative of ACM TOSEM, IEEE Transactions on Software Engineering, and ICSE aims to improve the quality and consistency of reviews of journal papers that extend conference publications, optimize the overall review effort of the community, and incentivize early publication of well-established results. The two submission cycles of the main software engineering conferences enhance the quality and consistency of the review process.

We can further improve the quality and consistency of the review processes, reduce the review bottleneck that stems from the dramatic increase of review demand, reduce the time-to-publication and time-to-presentation intervals of new ideas and results, and further improve both the conferences and journals of our community by sharing reviews and defining a shared review process in a seamless ecosystem that merges and enhances the Journal First and SCRE initiatives. We shall go beyond two review cycles within the same event and extend it to multiple venues with both conferences and journals. The community shall offer authors a common entry to submit their papers to conferences and journals and quickly present and publish valuable results and ideas. Authors shall be given the opportunity to indicate the target events, both conferences and journals, and the community shall organize the review process to share reviews and feedback.

Addressing the multiple and sometimes contradicting requirements of the complex conference–journal ecosystem within the framework of multiple societies and publishers and in a quickly evolving environment is a far-from-trivial task that we as a community can well address only with a joint effort among conferences and journals.