"Laws are like sausages. It’s better not to see them being made."If you want to know how talks for EclipseCon Europe are selected, take a deep breath, brace yourself and read on. I promise not to get into the really gory details.
The selection of the talks is done by a program committee of volunteers from the Eclipse community. You may have seen a few of us since we go fairly regularly to Eclipse events, but if you like to check for yourself, there is a page with all the PC members.
For EclipseCon Europe we have a fairly stable team with expert from all domains of the Eclipse ecosystem. Most of the members have been on the program committee for the last 3 to 6 conferences.
After the talk proposals are submitted (which is done at the EclipseCon web site just in case you still haven't send in your proposal) several things happen during a few weeks time.
First there is the is the early bird selection on July 17th, 2015. To give you an incentive to send in your proposal before the last weekend possible, we will choose five talks that will be accepted before the final deadline. The selection process is quite simple: every PC member sends her/his top five talks to me, I'll build a top ten list and the winners are chosen during a conference call on July 20th. Despite this afford roughly 80% of all proposals will be put in just before the final deadline on July 31st.
After all proposals are send in (hint: if it is August you are too late!) I will get the raw data from one of the webmasters and put it into a spreadsheet hosted at Google. This will invoke a lot of time and even more cursing, but eventually it will be done. The sheet is our working document for the next couple of weeks.
In the first step I will assign equal sized portions of the proposals to a pair of PC members each. They will check for talks aimed at the wrong conference, obvious sales pitches and also for some gems we don't want to miss (all in their opinion). The traditional -1, 0 and +1 are assigned. We will also try to contact speakers of proposals of which we think that some more work is needed. If you get a request from us and don't respond expect your talk to be dropped (Sorry, but if it is not worth your time, why should it be worth ours or that of the attendees).
On August 10th we will have our first selection conference call. We will briefly discuss the proposals which have gotten two -1 since they are probably to be dropped fairly quickly. Some more time is spend on proposals with one -1 since these need some investigation. At the end of this call about 10% of the proposals are gone.
Now follows the really time consuming part. Every PC member checks all of the remaining proposals and sets a score from 1 ("I think this is really bad") to 10 ("Won't miss it for the world"). This takes usually two long days of reading proposals.
After that we have list of proposals voted upon by every PC member. Okay, that is not completely right: PC members are not allowed to vote on proposals from their companies. Any other conflicts of interest are noted in the sheet and we can decide what to do with the vote.
An average of the scores and the standard deviation of all the grades is computed in the sheet for further analysis. I will set the threshold for acceptance so that just enough talks to fill all the slots are available.
In the calls on August 12th and August 14th we will discuss scores with a standard deviation above the "we agree" level, and the people with the highest and lowest scores will have to explain their choices. Then everyone has the option to adjust her/his vote.
During the final call on August 14th we will check if we think that all topics are reasonably well represented. We take the freedom to drop proposals slightly above our threshold to get a better balanced conference, if we think that a topic is missing from the schedule. This is usually the part where emotions rise high, but no blood was shed, yet.
On August 17th acceptance and rejection mails are send out. There is a very difficult task to be done before that happens. We have established a rule several years ago that every rejection note will get a comment why we had to reject the proposal. We try to give hints what was missing, but sometimes it's just a "Well, nothing was wrong, we just don't have enough slots." My apologies if this happens to you.
After the talks are selected the fun part begins where I try to schedule the talks to the best possible times and locations (I already did mention some swearing, right?). Luckily this torture ends on August 21st, when Anne Jacko expects a completed schedule (and you don't want to receive a call from Anne asking what's going on). On August 26th the schedule will be available on the web site and we can all look forward to another great conference! I will have spend about 20 days working on this, my fellow PC colleagues probably 5 days. It is time well spend since one learns a lot and gets to know a lot of people.
So this is how the talks at EclipseCon Europe are selected. I have left out the selection of keynotes (which is a very difficult task, but in the very capable hands of Ralph Müller) and the discussion about tutorials, which is mostly a small version of the general selection process.
If you are interested in being on the PC for EclipseCon Europe 2016, just be at this years conference and talk to one of the PC members. Maybe you will join the fun next year!
