Interview with Martin Lippert about the EclipseCon conference

1. I’m here at EclipseCon North America 2015 with Martin Lippert, who was the program chair for EclipseCon 2015 North America. Can you tell us how you got involved with organizing EclipseCon and when it all started?

You mean the EclipseCon this year, this EclipseCon? Because I don’t remember where my first EclipseCon was, it was maybe the second EclipseCon in history that I participated in! But my program chair works for Eclipse in North America started back last year already, I think it was even before EclipseCon Europe when Ian Skerrett from the Eclipse Foundation asked me to become program chair for North America, and then we started to create and set up the program committee, so find the right people that would like to put together as a team to work on the program for EclipseCon North America, because you can add the chair but it’s more putting everything together instead of deciding everything on my own. So there is a whole team of, I think we have 10 people on the program committee which is looking at the program and do all the work. But I think it started even before EclipseCon Europe in, that was October, so I think the call for papers went out shortly after that or around that time, or even shortly before EclipseCon Europe; and the early bird deadline was already I think beginning of November - so that was the time frame.

So we started pulling together the program committee, we discussed what we think should the program look like in general, so what tracks do we have, what tracks would we like to have in the program, what are the different topics, on which topics we would like to put some kind of emphasis, how would we rate those topics and how much space do we have in the end in the program because this year as you know it’s co-located with Foss4G which had kind of an influence in how many tracks do we have, how many talks can we accept, how should we split that out across different tracks. So we thought about those tracks and that was kind of the first steps for us as a program committee and for me as a chair.

2. And so once you decided what tracks were and how many presentations you could have in each of them, you can then go through the call for papers and try and start filling some of those slots.

Yes, we send out the call for papers, we announced it and we tried to find people to submit papers, so everybody was invited for that. From time to time I looked at the submission system which was completely open, so everybody could just watch the submission system, can propose a talk and can see what other people propose, so it’s quite open, transparent which is I think great for this kind of OpenSource community.

At the early bird deadline it was clear that we would like to accept I think 3 or 5 talks as some kind of early bird, say: “Hey, we would like to be nice to people who submit talks early” and then we selected a few of them. So we did that, we sat together as a program committee, looked over the submissions and say: “Ok, we really would love to pick those three”, I think we picked 3 talks as early birds. And after that I think it was 2 weeks left until the final deadline for the program, and after that really the heavy work started.

3. So what did you do then fill the rest of talks, did you force rank them against each other, did you let the track proposers manage them directly?

It really started with how we selected the program committee, because we knew that we would like to have certain tracks like a more platform oriented track and a modeling track and a cloud track. So we selected people from the program committee so that we kind of have experts for the different topics inside the committee, so that in the end we at least have maybe two people being really good in modeling and know what all these modeling proposals mean, because for example me I just have a rough feeling about modeling, the modeling topics but not real detailed knowledge, so I needed to have those guys on the program committee.

And after the final deadline we said ok, everybody goes through all the submissions, looks at all the submissions and we did some kind of early voting, so everybody puts a vote on every submission saying: “Ok, I think this should be in the program, this should not be in the program.” This is not really yes or no, so not a binary decision, was some kind of I think was between 0 and 5 stars, kind of judging the submissions a little bit. And then we set together as a group and say: “Ok, we would like to have, I think we have the space of 8, around 8 talks on average per track”, so we have I think 5 tracks, 8 slots per track so we said ok, what are the, what should we pick as the best talks, the best rated talks from the program committee for the cloud track, and then I think we pick maybe 5 or 6 of really high in the average rating from the program committee.

And we did that for each track so that’s how we filled the first slots for all the different tracks and then we kind of discussed the remaining slots because the next in the average rating, there were talks going up and we had some kind of discussions about, so we pick this talk or should be pick another talk, because sometimes we have many submissions and we have many good submissions, and it’s very, very hard to pick and have the two remaining slots there are 6 or 10 talks that haven’t accepted yet. They are really, really good and really would love to accept all of them, but you have to pick 2 or 3 and then there was some kind of lots of discussion going on among the program committee members about which one to pick, which one not to pick and those discussions, those were the hard parts, the hard work.

4. Of course this time we've had various days, like the iOT day and the CDT day and and so on, how did those come about and how were they organized?

The special days were organized upfront, so the special days was before the program committee started to work on the program, was: “Ok, we have the iOT day, we have the PolarSys day, we have the CDT day” and those days have their own track on one day and those days had their own program committee, so they had, if people submitted talks for the CDT day, those talks were judged and selected by the committee for the separate, totally separate committee for the CDT day for example.

So they selected their own program for their own days and we as a main or remaining program committee for the main conference, we had to pick all the other ones, so there was no, we do not have to take care about the CDT day, the PolarSys day, but there was of course some kind of discussion among the program committies as well because they had 2 or 3 submissions that people submitted for the modeling or for the Xtext day for example, where the Xtext day guys says: “Hey, those are great submissions but they do not really fit into the Xtext day, maybe they apply for the main program” so we took care of that as well after they selected their program.

5. So once you selected all of the talks that may fit in, how do you go about doing the scheduling, finding out when they are going to be on different day, estimating the size of the rooms needed and location?

We tried hard to find out what talks to put into the extended slots; there are a few slots in the program that are extended - the extended talk have 35 minutes I think. There are a few slots being an hour long, I think 3 or 4 slots a day, at least on Tuesday and Wednesday, not on Thursday, and that's always difficult. In the past when I was involved in the program committee for the EclipseCon Europe, we try and talk about let’s put the most famous and the most exciting talks into these extended slots to give them more room, and more time to talk about their topic because they are so exciting or so famous, but that turned out to be some kind of complicated and maybe a bad decision because it end up being, you look at the program and say: “I would like to attend every talk in this slot, in this extended talk slot, because all the exciting stuff it’s in there, so too many let’s say generally exciting topics in the same slot”.

So we learn from that and did it slightly different this time, we selected topics that we think, this topic is worth spending more time on because we really feel it is more detailed, with more content in a meaningful way. And we did not put too many talks that we think are interesting and exciting for the broader community in parallel with each other, so that was kind of a surrounding condition. And in the end there is always a bit luck and a bit kind of random what exactly to put where because you cannot put all the cloud talks in parallel, you cannot put all talks from the same speaker in parallel and things like that, so if take all that into account there is actually not much room left to really move things around. A little bit, and most of the work was actually done this year by Ian Skerrett so not even myself, I was very happy about that, because it's really hard to get the schedule right, and sometimes things change because people ask for "Oh I’m not there on Thursday, can I move my talk to Tuesday" and things like that, but that's usually possible, but the goal is always to avoid putting talks with similar topics in parallel for the same track in parallel and things like that.

And then you try to find out what topic is attractive to many people or to a smaller group of people to think about the room sizes and to roughly guess the room size, which sometimes works and sometimes goes totally wrong, but yes.

6. So, so far we've been back to this hotel many times for the EclipseCon Conference; presumably that means you have an instinctive feel of where the rooms should be or where the big rooms are.

Yes, but that was not done by myself. I think most of that stuff was done by Ian and the other organizers from the Eclipse Foundation, who are so familiar with all this layout, and the rooms and what’s possible. Anne Jacko does a lot of that stuff because she knows all the rooms and the sizes and get in contact with the hotel what’s possible and what’s not. Luckily I was not involved with them because it’s really complicated, it’s really, really hard, so my focus was really on selecting the right talks that I think or we as a program committee think, should be in the program, and let the organizational stuff be done by the people who really organize the conference on site which is Anne and Ian. Ian is an organizational, I mean the conference chair and Anne is I think the organizational chair, so they do those organizational issues and they do it great.

7. And so that means by the time you come to the Conference all of your work has being done and you can just sit back and enjoy the work.

Yes, I just sit back and enjoy, listen to talks, do my own talks maybe, if one of my talks was accepted - which is by the way not sure for people on the program committee submitting talks, they are not automatically being accepted, they are part of the selection process as all the other talks as well and there is some kind of secret rule on the program committee that you are of course not allowed to vote for your own talk or for talks from your company, for colleagues from your company, some kind of a secret rule inside the program committee. So I think there were people from program committee submitted talks and they were not being accepted for example, but I think I had a pretty good time at the conference, I just can sit back, relax, watch the talks and see if we picked the right ones and talk to people and see if they are happy with the program and what to change maybe for next year.

Alex: Sounds like a great position to be in.

Yes, it was a lot of fun, was a great pleasure, honor to do that for the conference.

Alex: Well it’s certainly been good so far; Martin Lippert, thank you very much!

Thank you!