Attempting conference changes: Why moving away from traditional publishers for proceedings is difficult

I’m currently the Steering Committee chair for the IEEE International Conference on eScience series.  (Note that I’m going to talk about IEEE in this blog, but the problem is really much more general; IEEE is just the particular example in this instance.) Our conference and its proceedings follows the computer science conference model, where full papers are submitted and peer-reviewed in advance of the conference, the accepted paper are published as the proceedings, and authors speak about their papers as the major part of the conference.

As an IEEE conference, we’ve published our proceedings through IEEE in the past, which has meant that attendees received a CD or USB stick with the full proceedings, and non-attendees can access the papers from the IEEE’s Xplore, or the IEEE Computer Society’s Digital Library, in both cases behind a paywall.

Now, we would like our proceedings to be open access, in part because we believe this will increase their individual impact and the conference’s impact and reputation.  A number of studies have shown that open access publications are cited more than non-open access publications, though other studies have shown little impact.

Open access & proceedings issues

For many attendees, particularly from outside the United States, their papers need to be published by a traditional publisher for them to count towards in the author’s institution’s evaluation process. This makes it difficult to move to a completely different, low-cost, fully open access model, such as figshare, zenodo, a university repository, etc., where we could keep our review model and the use one of these publishers to post the final accepted papers.

As a conference, we also take pride in our peer-review system, where 4 to 5 members of the program committee read and comment on each paper, and these reviews are used by the program chair(s) to decide which papers to accept, which leads to which authors attend the conference.

If we stick with a traditional publisher (e.g. IEEE, ACM, Springer), we or the authors would need to pay a large fee (~$700) for each paper to be fully open access.  We as a conference cannot afford to do this, and this would add 25-50% to the overall conference costs (registration & travel) for many authors.  Authors can generally post some version of their paper (either pre-review or post-review, depending on the specific publisher) to their own website or to their institution’s repository, but they cannot post this to a general repository for some period (usually 12-24 months), and we as a conference are not allowed to do this for the authors.

We could potentially move to a publisher such as PeerJ Computer Science or F1000Research.  However, PeerJ CS would not really reduce our per author cost, and both would make the review process quite different than it is now, reducing the control that the program chairs have over the papers that are associated with the conference.

Conclusions

Even though we would like to make changes in our proceedings to make them open access, it’s not clear how we should proceed, giving that we have an existing successful conference series with an existing review system we think works well.  None of the existing options seems to meet all of our goals.

If we were starting from scratch today, we might make a different choice, but making a change today seems to have a strong danger of losing a portion of our attendees, particularly from outside the US.

Any comments, or even better, suggestions, are welcome.

Acknowledgements

Other members of the IEEE eScience Conference Steering Committee and the organizers of the past few conferences were involved in many of the discussions I mention, and they have raised many of these concerns.  However, this blog represents my personal view on these issues.