SNAPL: A new kind of PL conference

The Summit on Advances in Programming Languages (SNAPL) is a new kind of PL conference, focused on big-picture questions rather than concrete technical results. The conference will be held for the first time in Asilomar, CA, from May 3 to 6, 2015.

The submission deadline is January 9, 2015 — if you have a long-term vision about where the field of PL should go, you ought to submit a paper.

Here we post an interview with Shriram Krishnamurthi, who is a professor at Brown University and one of the organizers of the conference. 1

What is SNAPL, and how is it different from existing PL conferences?

We want to be different from regular conferences in two ways: in content and in social setting.

In terms of content, a conventional conference paper is about a result: a concrete, cordoned-off, technical contribution. But a good deal of valuable scientific discourse is not about such results: progress on an ongoing, long-term research program; lessons from a completed project, especially focusing on mistakes; well-argued challenges to accepted ideas and methods; and so on. Such papers used to appear in conferences once upon a time, but our conferences have “matured” and evolved away from them. That leaves these kinds of contributions without any real home. Obviously, some places are thinking about work that isn’t conventional, such as POPL-OBT and Onward!. However, I think they’re quite different in character. OBT is specifically for short, snappy presentations of things that are strictly unconventional, while Onward!’s focus is on new ideas that are still short of evaluation, but expressed in the form of a traditional, full-length paper. You won’t find very many OBT or Onward! papers that tackle the example categories I’ve listed above.

Also, in terms of the actual meeting, conferences tend to be large and bustling, with the social pressures of finding friends and planning dinners and so on. This has its own value, but it’s worth contrasting to an event like a Dagstuhl seminar, which is small by design, self-contained, promotes interaction between all the participants, and has fluid boundaries between talks and other forms of interaction.

Is the goal of SNAPL, then, a Dagstuhl workshop based in North America?

We’re indeed big fans of the social environment of Dagstuhl. A key difference, however, is that the invitee list to a Dagstuhl event is determined by its small organization team. In contrast, we wanted an open event where everyone was welcome to submit, and selection would be based on the quality of submission.

Another difference from a Dagstuhl Seminar, specifically, is that our interest is very broad, whereas a Seminar is about a very specific topic. We are slightly closer in spirit to a Dagstuhl Perspectives workshop, but without the expectation of having representative oversight of the entire area, or the goal of setting an agenda for a community.

Note that none of these are criticisms of Dagstuhl. Every one of these types of events has its own value. Also, I might as well note that the SNAPL proceedings are going to be published by Dagstuhl!

So there won’t be a focus on certain themes?

No. The content has to relate to programming languages. That’s it. Beyond that it’s up to the PC (and those submitting papers!).

What would a good SNAPL paper look like?

Check out the Call for Papers!

What kind of review process will SNAPL have?

The review process is conventional. Our PC looks pretty conventional, too. However, they will be looking for papers that are intriguing, thought-provoking, and maybe the kind that would fall through the cracks of conventional conferences, with which they have plenty of experience. As chairs we’re trying to not dictate things; we selected people whose insight we value, and now we have to let them do their thing. Not that dictating is likely to work, anyway: they are, after all, a pretty opinionated bunch, so I don’t think they will listen to us very much!

What inspired the creation of SNAPL?

Our inspiration comes from many places. We’ve all enjoyed meetings such as Dagstuhl Seminars, IFIP Working Group meetings, regional PL events (NJPLS, NEPLS, etc.), Hot workshops, and more. SNAPL’s closest parallel is probably the database community’s CIDR series. In fact, the name is a mild tribute to that event. (We originally intended to call the event AAPL, but there’s a certain company with that stock ticker!)

While I have a chance to get this on the record, I should say one more thing about the origin. Since I’ve been one of the more visible people behind SNAPL, there may be a misconception that this was my idea. It was actually Ras Bodík‘s, so all the credit for proposing this should go to him. Working with Ras, Greg Morrisett, and Tom Ball on this is more fun than is permitted by the law in some places.

Tell us a little bit about the venue.

Asilomar shares most of the inaccessibility of other venues like Dagstuhl, which is of course both a bug and a feature. The compensation is being on the Pacific Coast. I hear it’s lovely. But I’ve never been there before; ask me again in May!

Do you already have plans for the next few iterations of SNAPL?

Not yet. I would like to see how this event goes. If it’s a success, great, we’ll start planning. If it’s a failure, well, we tried, but there’s no point dragging it out unless we can pinpoint what failed and people really want to take a shot at a revision.

Notes:

  1. My co-blogger, Mike Hicks, is on the SNAPL program committee.

Leave a Comment

Filed under Process, Research, Science

Leave a Reply