Experiences Organizing the SIGCOMM 2010 Poster Session

The following are notes written by Neil Spring about the 2010 Poster and Demo session. Please send updates or questions this page should answer to him. Running the poster and demo session is largely about avoiding a variety of potential disasters (knock wood), except that there's fun in coordinating the student research competition. Broad challenges in working with other chairs: (everyone did a fantastic job in the organization of sigcomm 2010)

Publication
The publication chair and Sheridan will think about the traditional TPC chairs and may leave you off important mail. Sheridan generates a prog_ck.csv file that includes all the "corrected" titles, author names, and affiliations that authors won't share with you. (From which the program gets generated.) Make sure you have access to this table.
Local arrangements
The main conference, meals, and workshops are likely to be thought of; good spaces for demos and posters less so. Hallways aren't ideal. Rooms that can be locked are good for demos since some take setup and have equipment. You'll need a meeting room for at least an hour for SRC. Power is a concern. Network cables, displays, display cables, tables, lots of stuff is needed.
General
There's lots to do, though poster decisions (almost necessarily) happen later than theconference and workshop program committee work. There may be a flurry of activity elsewhere (January through April) while poster chairs have to say "uhh, yeah, there's nothing going on" during conference calls.

Student Research Competition

ACM makes holding an SRC bizarrely easy. They want to know about the CFP's for their list, the list of students, and that's it. There don't appear to be limited resources to be allocated.

Authorship
SRC Rules state that the student entrant must be the sole author on the work. Various systems conferences have had multiple-author SRC work (still a single student as the leader of the work, but acknowledging the contributions of faculty advisors). Single-author student submissions appear to be a hardship, since it fails to acknowledge the guidance of faculty. The right answer seems to be to be clear about what the requirements will be early.
Undergraduate category
Getting undergraduates to submit posters to sigcomm probably requires specific publicity. We had only one (accepted) entry, which is disappointing because ACM provides three medals for each category.
MSR sponsorship
The MSR logo should appear on Poster-related pages.
Student membership
Surprise, many student entrants won't have their ACM student membership until after ACM contact them.
Judges
The registration chair should know who will attend. Everyone who was responsive to my invitation agreed. Ask for the list of participants shortly after the early-registration deadline because people may travel early for the conference.
Round 2 presentations
Need a room for a while.
Demos.
We treat the poster part of the demo as the SRC entry. We did not want to encourage someone who had a demo to show to submit a poster instead just to enter the SRC.
Ideas
Consider limiting the number of SRC entries so that judging requires a smaller, more easily managed committee. 20 entries were quite difficult. Consider excluding previously-published work at an early stage. (A poster would likely have this idea, but a sigcomm demo would not.)

Timeline (it is compressed):

December: Form PC. Check that CCR will take the best poster / demo abstracts so that can be included in the CFP.
January: CFP up.
Tell ACM about the SRC. Good to have the HotCRP site running or at least have the host name allocated.
May: Crunch time, also the end of the semester.
Submission after (full paper program) notification.
This permits authors of rejected sigcomm papers to submit a poster abstract without double-submitting.
Use HotCRP. On submission, track options for: is an SRC entry? (should also track undergrad or grad, though we did not) is a demo or a poster? Consider collecting technical requirements here.
Have three weeks for reviews... more if possible.
I used two rounds since we had 130 submissions and I thought it possible to reject some with two reviews (and keep better tabs on the committee).
One week for notification.
Comments / discussion after reviews is difficult to encourage. (we had less time; the PC chairs sought a deliberate review process and Sheridan wanted early camera ready deadlines)
Notification before travel grant deadline.
Post the list. Send the list of students to ACM and to the travel grant chairs. Decide on sessions if needed; SRC entrants should be early to accomodate the second round presentations. Send the list with schedule to Sheridan (TODO: attach script to extract from hotcrp)

During or after the conference

At the end of the conference, prepare the list of winners for a post like http://www.sigcomm.org/node/443 to announce the winners, and send it to the information services director.

Materials.

  • Monitors, should use sponsorship. Stands. Cables (vga).
  • Tables (dimensions).
  • Poster holders: The CFP says A0 portrait mode. It turns out that convention may be to have stands that are 3 ft by 4 ft landscape mode. Small issue. Try to determine what the venue will have before the CFP goes out.
  • Power consumption.

Demo "requirements"

  • People asked for three monitors (no).
  • 1kW (no).
  • Reliable Internet (good luck).
  • NAT and no-NAT addresses.
  • Wired ethernet.
  • 3G wireless reception. (fingers crossed)

Local printing. Shipping.

Customs was a concern for shipping demo equipment.

Numbers.

We had 128 submissions and 18 PC members: too few PC for that many submissions in the time between submission and notification. 2009 had 74 total submissions. Hooray for publicity chairs! Each submission had three reviews unless the scores in the first round rated (2,2), (2,1), (1,1), (1,x). Expect to accept ~10 posters and ~10 demos; more if the general chairs would like to be inclusive. The final numbers depend on the space. We had 20 and 20, split into two sessions of 10 and 10.