What are the challenges to publishing a paper when you have a huge number of collaborators?

Question

Following on from this question, I'm really curious to know about the logistics of large scale collaborations when it comes to publishing a paper that everyone has some contribution to.

To summarise, the recently published paper on the detection of gravitational waves had an author list with over 1000 names and 133 institutions. The sheer number of people involved would make publishing a single paper somewhat less straightforward than, say, a small research group based at a single institution.

So, to my questions:

• What were the logistical challenges of having over 1000 authors/collaborators when it came to publishing a single paper?

• How were these challenges overcome?

I want to make this question as specific and answerable as possible, so please edit this question as appropriate.

Answer 1

I have recently been the primary organizer and lead author on a large-scale collaboration, in my case an inter-laboratory study that took place over two years and with more than 100 contributing teams.

The paper has been accepted and is currently in press, and my experience so far has been actually a remarkably low amount of pain. The way in which we organized the study and the subsequent writeup had very clear roles, so there was never any question of having 600+ people trying to collaboratively write up a document. Instead, I led the writing with help from the other core authors; once we were satisfied with the draft, we circulated it to the whole community of authors for review, consent, and corrections, with a one-week deadline for responses. Lots of small corrections came back: we integrated them and submitted the paper as planned.

With all of those eyes on the paper, it is perhaps unsurprising that the reviewers requested only truly minor revisions, which meant there was no need to involve the entire community in the revision, just the core authors. A live-updated draft has always been available to the full community, however (Overleaf is great for this purpose), and additional small corrections have continued to trickle in, up to the very final revision.

In short, then, we simply short-circuited the challenge of many authors by using the project organization to divide into a small group that managed the actual writing of the paper and a much larger group that gave advice and consent. There was one very busy week while the bulk of the feedback was coming in, but other than that it has not been very difficult to manage the paper-writing process in the least, and it appears to have been a good experience for all involved.

Answer 2

An article on Retraction Watch today directly addresses this question. Here's a quote from When it takes a village to write a paper, what does it mean to be an author?:

[A]uthorship is determined by a set of criteria based on time in the collaboration and/or ‘service work’ – jobs like hardware upgrades, detector calibration, data-taking shifts, and the like, overseen by a hierarchy of institutional leads and, for the largest collaborations, national leads. People join the author list after meeting these criteria, and usually stay on until a certain amount of time (typically six months or one year) after they leave the collaboration. Authors are listed in alphabetical order; G. Aad is a prolific first author.

Everyone is an author on every paper, whether or not their specialized work contributed to that paper. Someone who worked on muon detection, for example, would be listed as an author even on papers that do not involve muons. There is no requirement that the individual authors have even read the paper, much less contributed to the writing.

Papers are actually written by a small group of authors or a committee, and internally reviewed by multiple committees. Individual authors typically have one opportunity to comment on the manuscript before submission to a journal. In some collaborations, individual comments are common; in others they are not encouraged. This is all governed by a collaboration governing document – a kind of membership agreement between the participating institutions. This is very different from authorship in a small group.