What Happens in the Planning and Execution of a Conference

(to be updated)

Consider a conference in June XXX9, Planning starts before June XXX7.

Conference Date - 2 years -- Before June XXX7

About 2 years (or 4 for ICCV) before a conference several people get together and decide they want to hold the conference. They then expand the orgainzing committee and look for reasonable places they would want to hold the conference in.

Conference Date - 2 years -- June XXX7

The presentation is made at a TC meeting at the approproate conference and the proposal is selected. Winning proposals are usually well organized, with specific hotel suggestions and specific current (non-negotiated) prices. Now the real work starts.

Conference Date - 18 Months -- January XXX8

The budget must be generated. See

IEEE CS Conference Information Page

for the spreadsheets. Also see

A gross level budget for CVPR

like conferences.

Conference Date - 18 months -- January XXX8

The initial start on forming the the program committee. One model that has worked is area chairs -- select about 20 people and they suggest 6-8 program committee members.


Conference Date - 12 months -- June XXX8

The preliminary call for papers should be available at the previous conference, with dates, locations, etc. The TMRF (Budget) should be finished and sent to IEEE.

Conference Date - 10 months -- August XXX8

Real call for papers, with electronic announcements, etc.

Conference Date - 7 months -- November XXX8

Paper deadline approaches. There is a flurry of mail and requests relating to the deadline. Is it real? When will it be officially extended? Of course it will be extended. What do you mean you cannot extend it? But I really want to work on the paper on January 1. I don't care that reviewers expect papers last week. There will be questions regarding the web site submission requirement. There will be complaints about any rules or limits you impose. There will be complaints about any rules or limits you do not impose. There will be complaints.


Conference Date - 5 to 6 months -- December XXX8

The paper deadline. Now the work starts. First, the initial assignment (under the current approach, CVPR has area chairs who do the assignment of the 30-50 papers in a given area, the program chairs assign to the area chairs). The papers are mailed out (Copy #1) to the area chairs who work for 1-2 weeks to assign them. When they notify the program chair handling the papers (i.e. fill in all the blanks on the www form). Copies 2, 3, and 4 are mailed to the reviewers at about D-5 months. (Before Christmas break.)

The planning for the program committee meeting should have started. A location near an airport, with minimal car rental expenses. A weekend -- the meeting easily takes 2 long brutal days.


Conference Date - 4.5 Months -- January/February XXX9

The committee should have made arrangements for the trip, hotel and air, but many will not. The flurry of messages about why people cannot make the meeting. Allow time in the schedule to move the meeting a week if a conference or major funding agency meeting gets moved to the chosen time. Note: In setting the deadline for reviews, reviewers tend to do them the last day (or even the day after the last day). It helps the review process if Area Chairs have sufficient time to go through the reviews and can ask for clarification from the original reviewers or seek additional input before the committee meeting. So set the review deadline before you absolutely have to have the review for the meeting and pressure reviewers to get them done by then. Then pressure area chairs to review the results and clean up any major variations in the results.


Conference Date - 4 Months -- February XXX9

Program committee meeting. Make the decisions. Depending on how quick you want to be, this can be before or after Presidents Day. (The holiday weekend may be good for a meeting, but it is bad for travel.)


Conference Date - 3 Months, 27 days

Rumors of the decisions get out and complaints start to come in about how many, which, which topic, etc. Deal with this eventuality at the meeting. There is always the question of whether the right decisions were made. Within 1 week of the meeting, post on the web site, and email the accept/reject notices. It is best for everyone involved to be out of the country at this time. Consider a trek in Nepal.


Conference Date - 3+ Months --- March XXX9

Complaints about rejections will arrive. There will be reasonable sounding arguments about papers that did not get 3 devestating reviews. Consider this possibility at the meeting.

Conference Date - 3+ Months -- March XXX9

This is also about when the final decisions are made on printing. Now CVPR and ICCV proceedings are only published electronically.

Conference Date - 3 Months -- March (12) XXX9

The first deadline from CS-Press on contents. See

IEEE Production Schedule

Immediately (so the boilerplate and planning should come before) after sending the list to IEEE, the preliminary program should be done. There are always delays here, so have a version ready immediately. Issues arise regarding wording, naming of sessions, etc.


Conference Date - 2 Months

This is when everything should be done. Registrations start to trickle in.

Example registration totals:

          C-N Weeks Member Non-Member Student Total
  8 Weeks 19 8 8 35
  7 Weeks 26 9 19 54
  6 Weeks _ _ _ 65
  5 Weeks _ _ _ 106
  4 Weeks 87 29 57 173
  3 Weeks, Friday 181 57 111 349
  3 Weeks, Monday 189 61 112 362
  End of pre registration rates
  2 Weeks 200 64 134 398
  1 Week 200 64 140 404
  0 Week 201 67 141 409
  Final for Santa Barbara 223 85 162 470
  Final for San Juan 203 85 120 412

Conference Date - 2 Months -- April XXX9

Final program. A/V requests for tutorials, workshops, posters, the main meeting. Exact rooms, exact menus and times for breaks and main social events. And Awards. And session chairs.

Conference Date - 1 Month -- May XXX9

The real details of session chairs. Finish the final program -- at most one month before the meeting for printing, etc. IEEE-CS will send summaries of registrations each week. At about 3 weeks the early registration time period ends and fees go up. These are the firmest numbers to use for planning. Walkup registrations are a moderate fraction of the total and happen even in remote sites (like San Juan).