Early-career researchers benefit tremendously from attending conferences. Alongside the obvious intellectual and academic benefits, conferences play a critical role in fostering a sense of community, helping us connect with our peers, and making the world of programming-languages research feel welcoming. Consequently, the format and frequency of conferences have a significant impact on early-career academics.
This past year, all four flagship SIGPLAN conferences (PLDI, OOPSLA, ICFP, and POPL) were virtual, and there has been ongoing discussion in our community about the future of PL conferences. Recently, another PL Perspectives post laid out some of the possible paths that we could take as the pandemic subsides. While deciding these next steps, we believe that the priorities and opinions of students and postdocs should be taken into consideration. This article lays out the perspective of five junior PL researchers from all over the world and at different stages of our early career paths. Some of us have also been heavily involved in organizing recent conferences.
What Early-Career Researchers Need
To develop conferences which serve early-career researchers, there are four main elements to consider:
- Community Building: Conferences are a major way for early-career researchers to find their place in the research community. Social interactions—in the form of gatherings, hallway conversations, networking, and more—can help. In-person interaction is a vital part of the conference experience, but successful experiments (using “experiment” in an informal sense) introduced at several SIGPLAN conferences this year show that some community-building can be done virtually.
- Research Dissemination: PhD students and postdocs must communicate their work to establish themselves as researchers. Often, a conference is the best chance for them to present their findings to other members of the community, which leads to new insights and collaborations. Since almost all research-oriented job interviews evaluate presentation skills, we must cultivate opportunities to present to new audiences.
- Accessibility and Inclusivity: Accessibility is a critical issue for everyone in the community, but it is especially important for junior members who are trying to get a foothold. We maintain that conferences must be accessible to those in different timezones, those with families, those without the means to travel, those with disabilities, and anyone else who is interested in being part of SIGPLAN and its events. However, accessibility needs to be balanced against the other needs that conferences must serve—accessibility to a conference is only worth as much as the conference itself.
- Climate Responsibility: Conference travel, especially when international, contributes to the warming of our planet. When considering options for conferences moving forward, early career researchers recognize we cannot ignore the carbon impact of conferences. The SIGPLAN community is already talking about climate change—we need that conversation to continue.
What We’ve Tried So Far
Physical conferences have been successful and popular for a reason—they are a time-tested and highly-effective method for promoting socialization and research dissemination. Spending the past year without physical conferences has underscored how important those in-person interactions are for sharing and discussing ideas with our peers and collaborators from across the world. Sadly, attending physical conferences can be difficult due to funding constraints, family constraints, and travel or visa constraints, leaving the benefits of physical conferences to be reaped only by those who are able to attend in person.
Virtual conferences have no such constraints. The online interaction mode means virtual conferences can benefit anyone, anywhere, anytime. In addition, many new and interesting events like Ask Me Anything (PLDI 2020) also become feasible in the virtual setting since speakers can more easily attend for just one day. Finally, as the only format that totally eliminates travel, virtual conferences are the most complete way to address conference-related climate impact. Unfortunately, the advantages of virtual conferences do come at the cost of significantly hindered socialization and research dissemination, which has the potential to harm the career prospects of junior researchers. Even though our community and conference organizers have done a remarkable job of making virtual conferences this year as successful as possible, it is safe to say that given the state of our current technologies, many of these physical conference experiences are almost impossible to reproduce in a virtual environment.
SIGPLAN has already noticed the drawbacks of existing conference formats. For instance, previous conferences used tools like sli.do and YouTube live streaming to make physical conferences more interactive for those not in attendance. Moreover, it is common practice for conferences to archive talks on YouTube, broadening research dissemination. These interventions have made great progress, but there is clearly more to be done.
The Best of Both Worlds?
This opportunity to reconsider the conference format will allow us to pick and choose the best ideas that the SIGPLAN community has to offer. Some will come from previous physical and virtual conferences while others are still being discussed. We have identified a few of these ideas that we think will become powerful tools for meeting our above criteria—the criteria that will make conferences successful for early career researchers.
Going forward, we support an approach that blends in-person interaction and virtual participation. This means making virtual attendees at our physical conferences first-class citizens. To this end, we propose that all conference talks be streamed as they happen, allowing both physical attendees and virtual participants to join in on the live Q&A. Moreover, pre-recorded videos should be an acceptable alternative to live talks, allowing authors to present even when they cannot physically attend.
For those around the world, conferences should support asynchronous interaction. Talks should be archived immediately so they can be watched on demand, and asynchronous Q&A should be available via a platform like Slack or Clowdr. This can all be part of a broader strategy of mirroring that increases global interactivity.
As conference technology improves, we should seek to create socialization opportunities between virtual and physical attendees. This can probably be accomplished most effectively by expanding formal socialization events. For instance, expanding José Calderón and Paulette Koronkevich’s trivia nights (one of the successful experiments into online socialization mentioned earlier) to allow both in-person and virtual team members would give attendees on either side of the divide a chance to interact.
One might wonder where our suggestions fit relative to the classifications given in the earlier PL Perspectives post. We see our solution as distinct from the four given proposals. While what we want is most closely related to the BBU and MIX options, neither of those meet our needs: “business as usual” does not expand accessibility enough, and MIX does not consider community building for virtual participants. Proponents of solutions like ours usually call it “hybrid,” but that term means different things to different people. Because the hybridization we are suggesting here is closely related to what past physical SIGPLAN conferences were already doing, it fits with many of the ideas behind BBU. However, we believe that expanding our use of virtualization technology to increase accessibility is vital.
Even if we adopt all of the suggestions we advocate for here, conferences will continue to evolve—it is critical that this evolution happen in an intentional way. With this in mind, we should add one more thing that early-career researchers need from the community: improved data collection.
In the 2020 ICFP report, Stephanie Weirich and Benjamin Pierce note that 1,155 people signed up for ICFP 2020; that is almost double the 586 attendees from 2019. However, according to the same report, the ICFP keynotes only had a maximum of 287 concurrent viewers, and the most-viewed session got just ~570 unique views with ~150 of those being concurrent. Jan Vitek also noted in a comment on an earlier blog post that he would have liked someone in charge of data collection at SPLASH 2020 so he could better understand who participated. These organizers worked hard at collecting their data, but the fact of the matter is that it is difficult to draw clean conclusions about inclusivity from these statistics. More broadly, conference surveys and data collection initiatives are not unified across SIGPLAN, so they often ask different questions and get incompatible answers. With this in mind, we propose that SIGPLAN form a Committee on Conference Data.
The committee would be made up of: one organizing-committee representative from each of the flagship SIGPLAN conferences, one early career representative, and, crucially, a professional data collection specialist hired by SIGPLAN. The group would identify and collect key data that is pertinent to conference organization, especially with respect to physical versus virtual conference formats. The committee would make data-driven recommendations to SIGPLAN organizers based on the collected data and guided by core tenets such as community building, inclusivity, research dissemination, and climate responsibility.
We realize that this is not a small request, but we are confident that it is both necessary and achievable. If the committee were to form by May 1, 2021, it would be able to start collecting data at PLDI 2021 and continue through the next two years, providing enormous clarity for SIGPLAN organizers at a time when so much is unclear.
Joseph W. Cutler is a final-year undergraduate at Wesleyan University, where he works with Dan Licata. He will proceed to a PhD program in Fall 2021.
Harry Goldstein is a PhD student at the University of Pennsylvania working with Benjamin Pierce. He was on the organizing committees for ICFP 2020 and POPL 2021, working on A/V coordination and social events respectively.
Andrew K. Hirsch is a Postdoctoral Researcher at the Max Planck Institute for Software Systems in Saarbrücken, Germany, where he works with Deepak Garg. His research brings ideas from the theory of programming languages to security.
Jaemin Hong is a PhD student at KAIST, where he works with Sukyoung Ryu. He was an SV co-chair of ICFP 2020 and will work as an SV co-chair at ICFP 2021.
Chandrakana Nandi is a PhD student at the University of Washington, where she is advised by Zachary Tatlock and Dan Grossman. Her research is at the intersection of PL and computational geometry.
Disclaimer: These posts are written by individual contributors to share their thoughts on the SIGPLAN blog for the benefit of the community. Any views or opinions represented in this blog are personal, belong solely to the blog author and do not represent those of ACM SIGPLAN or its parent organization, ACM.