Innovating an Innovative Conference on Innovation
Innovation in the theory of computing
Adolph Ochs was an American newspaper publisher, and is credited with creating the famous motto of the New York Times: “All the news that’s fit to print.”
Today I have been asked by Shafi Goldwasser to announce a new conference, well the third version—with a new name—of a conference on innovation in theory.
As a general rule we do not make announcements here, since there are better places to do that—places we all go to see the latest news, such as here. But this is the link to the conference, which is now called ITCS (Innovations in Theory of Computer Science). The reason for the name change from ICS to ITCS is according to Shafi:
it’s now an ACM sponsored conference and they didn’t want us to use that name.
As I said, Ken and I do not usually do announcements, but I feel that there are some interesting issues that this conference raises. Reversing Ochs’ famous motto, I want to suggest that we consider a new motto:
All the news that fits we print.
This is a much used old joke—but relevant for today’s discussion. We will suggest that there be some venues in theory that implement a different take on “survival of the fittest”: let all the ideas have free range on the safari first, then see which survive.
It seems to me that the talks at the last two I[T]CS conferences were by terrific people, on neat topics, and I wish I could have been there. But it also seems that they were not all that different from talks at STOC or FOCS or your favorite theory conference. This is not a complaint, yet it does seem like a lost opportunity. If the goal of the conference is to promote new ideas, new directions, and new research programs, then I think it did as well as our usual conferences. But not vastly better. Indeed I have heard comments from attendees that echo this opinion—a great collection of talks, but
I have a suggestion on how to organize the conference. The neat thing about having no formal connection to the conference is I can make suggestions, without any constraints, without any worry that they might actually be followed.
The conference should be semi-parallel: some talks are non-parallel and most are parallel.
Non-Parallel Talks: These should be invited special talks by leaders in the community. I am always interested to hear what X, Y, and Z have to say. Hopefully they will lean toward innovation—but that is up to them.
Parallel Talks: These talks will be all the submitted papers. All. Not most, nor half. All. No papers will be rejected. None. The only control I would suggest is that a submitted paper must also be public: it must be posted to the “http://arxiv.org/.” This forces the authors to stand behind their work at least to the extent that they are willing to have it archived forever.
What is the advantage of accepting all? Clearly not all papers will be great, nor will all be correct, nor will all be on-topic. But researchers with anything to say will have a chance to say it. Crazy ideas will be there, crazy ideas that are just silly. Along with crazy ideas that years from now will be viewed as seminal. Ideas that may change the theory landscape. If a talk is out there, then it may attract few listeners. That is the risk the author of the paper takes.
The Association of Symbolic Logic has used exactly this format for their conferences: some invited talks by leaders and other talks which are completely unrefereed. They survive having no barrier. I suspect that many of the talks are not landmarks, but that is to be expected. Few talks are. I once attended a talk at one of their meetings on:
An Inconsistent Number.
The author had previously claimed that set theory was inconsistent, then that arithmetic was inconsistent, and finally that even primitive recession functions were inconsistent. It was wrong, but so what. Funny it is the only talk I recall from the meeting. Oh well.
What do you think? How should a conference on innovation be organized?