I am a secretary for the American Geophysical Union (AGU). Specifically, a secretary for the Magnetospheric Physics section (SM). When I agreed to run for this post, I didn’t really realize what it entailed. I though I would write a bit about how the Fall AGU meeting in San Francisco is planned. The Fall Meeting is growing by leaps and bounds every year, so it is becoming more and more difficult to plan it all out. For example, there were over 21,000 abstracts submitted to the meeting.
The first part of this post is aimed at people who don’t know much about AGU in general, while the later part is aimed at people who might actually want to know how the meeting is really organized. In gory detail.
For the lay person, the Fall Meeting is a week long meeting in which there are both poster presentations (given on ~3 foot by ~4 foot poster boards) and oral presentations with lengths spanning from 12 to 30 minutes (the most common is 15 minutes). There are over 20,000 people expected to attend this year. Oral sessions are 120 minutes long (so, nominally 8 x 15 minute talks), and there are 4 of them in a day (running from 8 AM to 6 PM with breaks and lunch included). There are 45 rooms that are all holding oral sessions at the same time. Typically, the oral session rooms hold about 300 or so people, but attendance varies wildly from almost empty rooms to standing room only (depending on time, day, speaker, topic, etc). Friday afternoon is the worst time to give a talk at AGU, since many people have left by that time. Monday morning is a bad time also. Giving a talk at 8 AM almost any day is bad, but something like Thursday at 8 AM is really bad, since everyone is tired and shows up late.
The poster hall can hold about 2600 posters per day. It is hard to imagine what 2600 posters looks like – think of it this way: 2600 x 4 ft = 10,400 ft, which is 2 MILES. With so many abstracts submitted (>21,000), there are no spare rooms and no spare poster spaces. Everything is very tight. Giving a poster on Friday afternoon is like giving a poster on the moon. You will be there with only the other people giving posters. No one wants a poster Friday afternoon.
So how does all of this get organized? It is a multi-step process that starts at the beginning of the summer.
1. Sessions Proposals are submitted. Sessions are those 120 minute slots. If you have an idea for a topic that you would like to see presentations on, you can write a short proposal (maybe a paragraph to a page) saying what the topic is. You should team up with 1-3 people who might also enjoy seeing such a session. Be careful, though – if you are “awarded” the session, you actually have to do work. And if the topic is popular, it might be a fair bit of work!
2. Once the session proposals are all submitted, the secretaries for the different disciplines get to look them over and judge them. We are supposed to see if they are too specific or too general or too outside of the field or whatever. Most of the time, we leave them alone and accept them all. If the secretary would like to merge some sessions (since they are very similar) or reject some (since they think that not enough people will submit to the session), it is a good time to do this. As a secretary, you don’t want to be too restrictive, but at the same time, it is better to merge sessions at this point if they are quite similar to each other. Another thing that could possibly be done at this time, which I don’t know if anyone actually does, is make up a session that is blatantly absent from the list of proposals. For example, this year there were no session proposals for reconnection physics for the magnetosphere. There were ~25 abstracts submitted to the general magnetosphere contributions focused on reconnection, which is enough for its own oral and poster session. As secretary, you can make a session proposal.
3. There is a meeting at AGU in Washington DC in which there is some discussion about something. I have no idea what the purpose of this meeting is, since I have never gone. Only one person from the disciple is allowed to go (i.e., only one person from space physics) , and I was never the one who had to go. At this meeting I think that they decide officially on the session proposals and talk about the schedule of events leading to the fall meeting.
4. The people who submitted sessions are informed of the decisions. The secretaries don’t do this officially – AGU does. I will now call these people the session chairs, even though this is not technically true.
5. The session chairs decide on who they should invite to submit an abstract. Typically invited people are given oral presentations (talks) of some extended length (20 minutes is common). There are some issues with this: (a) you can only invite 4 people; (b) if you invite people, but no one else submits an abstract to your session, you might only get a poster session, so your invited people will have posters and will be quite cranky at you; (c) you might only get a half session if not enough people submit, so the only talks you can have are your invited speakers; and (d) you have to trust that the people who you invited will actually submit an abstract. I ran into this issue this year – invited person didn’t submit an abstract (oops). The session chairs have to enter the invited speaker’s information into the system. They are then invited by AGU to submit an abstract.
6. The abstract submission site opens. Chaos reigns. Very detail oriented people submit abstracts early, while the rest of the community waits until the last day or two before the deadline. In theory, a good secretary should send out reminder e-mails and push sessions. There are typically a couple of e-mails that go out through the SPA newsletter that details what the session are. At 23:59 UTC on some date set in stone, the deadline occurs. The website for abstract submission stops working. After this point, AGU will not take any abstracts. None.
7. Now the fun begins! AGU takes a couple of days and they tally the total number of abstracts submitted and the total number submitted to each session. They allocate a certain amount of oral sessions to each discipline. They give the secretaries the spreadsheet of all of the numbers and the allocation numbers. Access to all of the abstracts opens up, so the secretaries can see all of them. This year there were over 21,000 abstracts. I personally had about 850, which is a relatively small amount. The smallest is probably around 200, while the largest is probably on the order of 3000.
8. The secretaries now have to allocate oral slots to the session chairs. This year, if you took the number of abstracts, and the number of oral slots and divided, it came out to 24 (which is the number that AGU aims for every year now). This means that if 24 abstracts were submitted to a session, then they would be give one oral slot and one poster slot. Ideally, they would then have 8 talks of 15 minutes each in the oral session and 16 posters (i.e., 1/3 orals and 2/3 posters). 24 abstracts is ideal. 48 is ideal. 72 is ideal. etc. The problem arises when you get sessions that have 12 abstracts submitted.
There is a general pool of abstracts – people submit to the general session when they don’t find a specific session that they think that their abstract will fit in. A common practice for secretaries to do is to go through the list of general session abstracts (I had about 125 this year), and start allocating them to other sessions where they might possibly fit. Sometimes this is stretching things a bit, but it is better to “bulk up” the smaller sessions. Also, secretaries can merge sessions at this time too – take two sessions that are similar to each other and combine them. It was easier to do this BEFORE abstracts were submitted, but it can be done at this stage too.
The somewhat silly thing is that you can’t actually MOVE any abstracts from one bin to another bin at this time. You can only allocate numbers. This is an extremely frustrating aspect of the AGU system. As a secretary, you allocate the numbers and tell the session chairs that you are going to move abstracts or combine sessions, etc, but you can’t actually DO it right now.
Another very common practice is to award 1/2 sessions. If you find two sessions that only have ~12-16 abstracts submitted, but you can’t combine them, you can still put them together in the have 120 minute slot, with one session taking the first 60 minutes and the other getting the second 60 minutes. Session chairs HATE this. They complain bitterly about it, since they probably invited 4 people, so all of the (invited) talks will have to be 15 minutes, and they can’t have any contributed talks at all. But, there is nothing you can really do about this.
If a session chair gets on the order of 8 abstracts submitted, it is very bad news. They probably should only get a poster session or should get merged with another session. Very sad.
Also, at this time, secretaries can look through other discipline’s abstracts and see if they can get any of these, or give any of theirs away. This takes a LOT of time to do. But, if you are a good secretary (which I am not), you would do it.
By the way, I should mention that AGU gives you a few days to actually do this. It is an extremely tight schedule, where you, in essence, drop everything and work on AGU stuff for at least a day or two. But don’t worry, it gets worse. Much, much worse.
9. Once the allocations are set up, AGU creates the actual sessions. They are not ordered or anything at all, it is just a massive listing or oral and poster sessions.
10. On a Friday night in late August the website opens where the session chairs can start moving abstracts into their sessions. At the same time, the secretaries are given access to this site. All of the bargaining and such that occurred in the last step has to be implemented. Typically, this means that the secretaries HAVE to log in ASAP and start moving abstracts, before session chairs “forget” about the deals that they made to allow them to have a full session or three sessions instead of two sessions. Often the session chairs are very good and they actually do what you ask them to do. Sometimes they take all of the abstracts that you give them and simply put them into the poster sessions (This angers the secretaries. You won’t like an angry secretary. Remember the ultimate weapon – Friday afternoon session!)
This year, AGU gave us five days (two of which were over the weekend) to do all of the moving of abstracts. And one of the days overlapped with the meeting that I will be discussing next. This is really, really, really bad. There were many extremely angry secretaries. I, personally, worked all weekend and a full day, with some mopping up pieces on the 4th day. Many secretaries had issues because all of the session chairs were on vacation during this time. It is almost impossible for secretaries to figure out who should get a talk or a poster for a given session, since they don’t know what the session chair actually wants.
11. At the very end of August, all of the secretaries get together in Washington DC at AGU headquarters to have a meeting in which the Fall AGU meeting is put together. In theory, we have all of the populated (oral and poster) sessions in a pile, and we just have to schedule them in what order we would like. There are 45 rooms available for talks at any one time and a gigantic poster hall that can hold a maximum of about 2600 posters (or so AGU told us – it turns out if you take the number of posters and divide it by 5 days, you come up with 2652 per day, which is slightly more than 2600 – oops. More on this in a bit.)
When you arrive, there is a wall of note cards that have the allocations for each oral session line up by room and time. These are general allocations, not allocations for session. For example, Magnetospheric Physics was given room 308 all day, every day for the week. On Monday and Tuesday, we also got room 306. This means that on Wednesday, Thursday and Friday, there would be no overlapping sessions, which is great! On Monday and Tuesday, there would be a total of 8 parallel sessions. You get note cards for all the oral sessions, and you have to play with them until you come up with a schedule that you are happy with. And then there are the complications. If you were only working by yourself, this would only take about two hours to do.
What I do is open a spreadsheet and put the rows as times and the columns as rooms, then start playing with the ordering. It is somewhat important to do this on a spreadsheet, since the cards are too hard to work with and can get easily out of place. The sessions can change quite dramatically over the course of 2 days, too.
One complication is that there are other disciplines that are similar to yours. They schedule sessions independently also. Then you compare and move things. Compare and move. Compare to a different person. Move. Compare to the first person. Move again. Cry a little. Move. Frustration sets in. Give up.
Then someone comes up to you can want to swap times or something since they just CAN’T make their sessions work without a miracle. And you are the miracle. This time I was give a HUGE break – someone was so desperate to get a session the traded me an oral slot on Thursday for 3 oral slots – two of which were Friday afternoon (ouch). I talked to some session chairs about it (two of which had 1/2 sessions, and I offered them full sessions – but on Friday afternoon). They leaped at the offer. So, I made the deal, and essentially had to start the scheduling over again, given that I couldn’t move session X, Y or Z because I had committed to others that I would leave them there. More crying.
Now, once the oral sessions are “done”, you have to schedule posters. There are massive complications with posters. Ideally, you would like the posters to be AFTER the oral sessions, so people can advertise posters that have something to do with their talks. Second, no one wants a poster on Monday morning or Friday afternoon. They just don’t. Third, because of this, everyone schedules posters on Tuesday-Thursday, which means that there is a big bulge on these days, which AGU won’t allow, so you end up having to move posters to Monday and Friday. Therefore, what you typically do is oral slots on Monday afternoon get posters Monday morning. Monday morning orals get posters Monday afternoon. Tuesday AM orals can get posters on Monday also. You can’t have orals and posters at the same time, so you have to shuffle the poster sessions around until it all works.
This year I had a very hard time because I swapped the session in the middle of the day Thursday for three sessions on Friday, which means that there is not a lot going on on Thursday, but there is a lot going on on Friday. Thursday I have all of the same topic oral sessions – so I couldn’t put their posters on Thursday. Their orals go until lunch on Friday. So, the only two options that I really had were putting them on Wednesday or in the afternoon on Friday. One AGU hates and one the session chairs hate (kiss of death). There are all sorts of these decisions that have to be made. Every aspect of scheduling the fall meeting is a huge compromise.
Once all of the orals and poster sessions are scheduled in your head and on your spreadsheet, you have to make sure that the note cards are in the right place and have all of the right information on them. This is very important for AGU, since, if the system crashes and the lose everything, they have a huge paper trail that will allow then to recreate the meeting.
Independent of the note cards (and spreadsheet), you do everything on the website too. You have to do all of the creating new sessions, merging sessions, moving talks, etc. on their site. AGU (actually ScholarOne) just rolled out new software this year to do the scheduling, which was somewhat of a nightmare. We were using a beta version. It has to be tested sometime, obviously, but it was somewhat painful.
Then you have to make sure all of the names of the sessions are correct. For example, if you have 2 oral sessions and one poster session, they have to be named:
Cool Session I
Cool Session II
Cool Session III Poster
or, if the posters are before the orals:
Cool Session I Poster
Cool Session II
Cool Session III
THIS IS NO JOKE! It is extremely important that you change all of the session names to be like this or else you can never leave AGU headquarters. I kid you not! And, the names have to be consistent in the computer and on the cards. So, once you have changed all of the names on the computer, you have to go gather all of your cards from the wall of cards, change all of the names, and put them back up. At this point, all of the secretaries are crying and mumbling things like “brain dead monkeys could do this job…” There is much swearing.
This is a picture of the row of cards from last year’s meeting. This year they added two more boards.
A few more issues that I didn’t discuss above, at all: (a) Each talk has to have a time associated with it. There are 120 minutes per oral session, so if you have 8 talks, then each should get 15 minutes. If there is an invited talk, then you can do something like 22 minutes for the invited talk and 14 minutes for 7 other talks. Or if you have 2 invited talks, 18 minutes each plus 6 talks for 14 minutes. You have to play some math games to make them all work out, so the session is very close to 120 minutes, but you have maximized the number of talks and the time for each person. Session chairs should do this, but they sometimes don’t. Or they don’t do the math right, so the secretaries have to go in and change everything. I have heard of session chairs just going in and making the session 135 minutes long, and filling up this amount of time. That is a big no-no, and the secretary will be forced to change it back and delete talks. Also, I have heard of someone giving out something like 17 x 7 minute talks. This is just insane. Not too many people can give a very good 7 minute talk. (b) Talks and posters should be in some meaningful order. This is somewhat easy with an oral session or two, but is somewhat complicated with 100 posters. Some people try to do this, others don’t bother. (c) You have to make sure to have people who are actually going to convene the session (i.e., really chair the session). In theory, all of the session chairs should be conveners of their sessions, but in practice, this sometimes does not happen (especially if there are MANY sessions on the same topic). Secretaries often have to find conveners. (d) Abstracts that have fallen through the cracks have to be put somewhere. Technically, you should look through every abstract and see if it belongs or not. If not, you can put it in the “Exchange Bin”, in which case someone else can pick it up. You are supposed to troll through the Exchange Bin often during the meeting to pick up any abstracts that have been abandoned. The strangest thing that I saw this year was an invited talk in the Exchange Bin. How does that happen??? Well, someone dissolved a session because it got too few abstracts, so they put the invited talk in the Exchange Bin. Very odd.
Once all of this is done (for all of the different disciplines), AGU is ready to see if there are conflicts. This happens if someone is scheduled to chair two sessions at the same time, give a poster while they are supposed to be giving a talk, giving two posters at the same time, giving two talks at the same time or chairing a session and giving a talk in a different session at the same time. All of these conflicts suck. Some are easy to fix, while others are extremely difficult. For example, if someone is scheduled to give two talks at the same time, you can reorder one or both of the sessions (session chair could be involved or not, since this is something you have to fix pretty quickly). If they are chairing two sessions at the same time, you just find a new chair. If they are giving two posters, you could swap poster sessions and hope that it doesn’t raise too many new conflicts, or (more commonly) you say “tough luck”, and let them give two posters at the same time. Once you have gone through all of the drama to get this far, it is quite difficult to deal with these types of issues. Emotionally, you are pretty much cooked by this point. You just want to be done. This is the point in which detail oriented people shine, while 80% good enough people (like me) just throw in the towel and give up.
After all of the conflicts are resolved, AGU throws a spreadsheet on the projector and shows that we have too few posters on Monday and Friday and WAY too many on Tuesday, Wednesday and Thursday. They say things like, “No one is leaving until we are below 2675 posters every day.” Everyone cries at this point. No one is willing to move anything. Typically there is a stand off for about an hour, in which nothing really happens, and then the the will of the people start to crumble and poster sessions start to move. This, obviously, is a huge challenge, since everything is balanced perfectly at this point. The real problem is that you can’t just move poster sessions a day or two away from the corresponding oral sessions. Who would want to be in an oral session on Tuesday and have the posters for that on Friday? No one. Therefore, the whole thing starts to cascade. I personally think of it as a house of cards – you try to see which cards you can take from one place and put into another place without causing the whole thing to fail.
Typically there are roughly three rounds of moving poster sessions. Each time, AGU reruns the numbers, throws a spreadsheet up on the screen and everyone moans. Finally, the numbers drop down below the goals for each day and the killing of the trees begins.
By this, I mean that once your schedule is set in STONE, you have to print out the entire program for your discipline (posters and orals separate). For me, this was about 40ish pages for the orals and 80ish pages for the posters. For others, it was a whole ream of paper. The AGU staff them takes these printouts and compares them to the note cards that you have on the wall. They go through EVERYTHING. It takes them a good 20-30 minutes for each. They come back with details that you have missed, and then once you correct these details, you are free to leave the meeting. Glory be!
Here is a picture of my finished cards for Magnetospheric Physics (the green SM cards) from this year (with my phone – sorry that the quality sucks!). The rows are times, while the columns are rooms. Actually, the rooms are two columns wide, with two session pairs (two morning and just below that, two afternoon). You can see that I have double rooms on Monday and Tuesday. Wednesday I have a single room. Thursday I have a single room, but one time-slot is missing (this is the one that I traded). On Friday I have two rooms almost all day. All of the times in 306 (3 slots) I traded for.
12. All of the abstracts that have been scheduled are accepted, and AGU sends an automated message out to everyone to tell them when their presentation is.
13. At this point, I think that the secretaries job is pretty much done. There is not much you can do about the AGU meeting after this, since everything is in concrete. AGU organizes things like housing and registration and all sorts of other things.
Hope you now have a better understanding of what goes into making the Fall AGU meeting!
I have to thank a few people for helping me through this ordeal: Larisa and Ian for doing this for SA and SH; Linda and Alin and the other people at our table who listened to me cry and complain for 3 days while they pretty much did the same; and Joanna and Melissa from AGU, who have to organize all of us – they are unbelievably patient with all of us through this whole ordeal. The dude from ScholarOne who gave me superpowers for an hour so I could help someone else schedule their session (the power!). Obviously, I should thank all of the session chairs. I am sure that there are many other people who I should thank too, but I am exhausted. Sorry!
What I would change if I were king of AGU:
1. I would give all of the secretaries a manual of what we are actually supposed to do when, why and how.
2. I would allow the secretaries more time to do things, instead of having very short fuses on everything.
3. I would start the 3 day meeting from hell a couple of days earlier on-line. If the secretaries knew how many sessions per day they had allocated, they could start very early.
4. I would have some sort of scheduling software that would make it easy to move sessions around and ply with the schedule, instead of having everyone do this in excel and then moving things. Maybe the scheduling software could print cards with the correct names on them too!
5. The conflict software should be running continuously, looking for potential conflicts and reporting them on the fly.
6. There needs to be more room in the program for slop. There is no room right now. In talking with AGU, they expect continued growth for the next N years, so they need to have buffers in place already. There are no buffers. It shouldn’t matter whether there are 2725 posters or 2650 posters on a given day. I would have slop!
7. I would consult with the community before dramatic changes in schedules are made!
In conclusion, the Fall AGU meeting is very big. It is very complicated to schedule. AGU does many things right in the scheduling of the meeting. Somethings could definitely be improved. The community should be aware that the scheduling of the sessions is compromise after compromise. It is difficult to get everything exactly how everyone wants it, or how it SHOULD be.
Lastly, the AGU secretaries are not paid. They work their butts off to make the meeting happen. Please, thank them!