I have never witnessed any improper conflict of interest, and I don’t think there are any in the community which would try to subvert PyCon… but…

Conflict of interest is a difficult problem, and there are different aspects of it to deal with.

Foremost we need to trust our reviewers. What I mean by this is that there is no way to be 100% sure there are no conflicts of interest at play. There are no automated or software solutions to solve this problem. It is inherently a human based issue. As such draconian procedures will only cause problems. Most people are more sensitive to their own perceived conflicts of interest; that is we judge ourselves more harshly than others would on average. We need to trust that our reviewers will ‘do the right thing’ as it is impossible to prevent it otherwise.

With that said there are checks and balances.

There are two primary forms of acting on conflict of interest:

A. constructive (author is a known associate, etc)

This is the most common, as the python community is not that large, especially with domain topics like web2py or jython. This is not a bad thing as we want those domain experts reviewing these talks. Also information like ‘this author is a friend of mine. he gave the talk at the local users group and it was well received.’ is information crucial to the program committee. Here we ask that reviewers detail their conflicts of interest in the review.

B. Subversive (collusion and sabotage)

This includes the scenario you bring up. These are situations where a person receives some sort of benefit from seeing a talk accepted or declined. Here there are three checks to try to detect when this happens.

1. At least 3 reviewers

Every proposal must have at least 3 reviews. It is much harder to improperly review a proposal when at least two others are also reviewing the same talk. Many of these reviewers are prominent in the python community. Passive pier pressure can be a powerful motivator.

2. Reviews do not determine acceptance

We do not use a computed score when writing reviews. The mark given is just to help break down the talks into interesting groups. As such it is impossible for someone to kill off a proposal. The PC as a whole discusses the proposals by breaking them down into successive groups. Part of that is breaking them down by overlapping and intersecting topics. It should be obvious if a champion speaks up against a talk then turns out to be the author of of a talk in that same group.

3. Authors get to see the reviews and provide feedback

Before the accepted talks are announced to the general public, the proposal authors get to see the written reviews and provide feedback. Many times they are in the best position to identify when there is a subversive conflict of interest.

At the end of the day you need to trust the community.

I am curious about one point which you did not address: How do you handle conflicts of interest? If Bob, who volunteered as a reviewer, submits a talk, how do you ensure that he is fair in his assessment’s of a talk by Jane on the same topic? Perhaps Bob’s proposal was written after he read Jane’s, etc. Just curious…