Before we go into the learning bit, let’s just revisit what a version of record is. Wikipedia describes it as “the fully copyedited, typeset and formatted copy of a manuscript as published” (with two references). Basically, in the whole scheme of research output, it is a release. It is a tagged version of the output, allowing people to discuss that version specifically, so that we do not run into endless “oh, but I meant version manuscript_rewrite_V2_AE_MB_Fixed.docx”. Really, publishing is not unique at all and publishers are doing it wrong.

So, with “version of record” defined, why do we have only one in publishing?

There is absolutely no reason not to have multiple versions of the same narrative, as long as they are clearly tagged. Open Science has been doing this for two decades, but publishers have been slacking. Retractions are updated versions of the same article, as are corrections, corrigendum, and errata. It is not that publishers do not know how to do it. Hesitently, they are accepting that preprints exist, but publishers tend to frame that as inferior versions. There was a paper earlier this month that looked into how much the versions are really different, and when I find it again, I will add the link.

Of course, money, control, power likely have a role here. And historic reasons too, I guess. After all, when you have to print a journal issue and send them by horse carriage to universities around the world, making updates is indeed not trivial.

Twitter or X (or Mastodon or Bluesky)

But as an openscientist, I have the urge to keep research output relevant. We do this for data, we do this for community standards, and we do this for research code. Routinely. Again, for decades. Must open access not learn from open science here?

I asked that recently on Mastodon. Specifically, should the Ten simple rules for getting started on Twitter as a scientist article (doi:10.1371/journal.pcbi.1007513) not be updated? Looking at the number of scientific papers that discuss social media in scientific communication, ten tips sound to me to be timeless. And I was interested in why or why-not the paper should be updated. Content-wise, a trivial update would be to update the name to X, which is the name of what was formerly known as Twitter. But then, updating the paper to replace Twitter by Bluesky or Mastodon would not be that much work either.

The discussion brought up various aspects of this question (and hereby thanks to all who joined the discussion!). Is it worth it? Is it legal? Should it be an update, or just a new paper? Who should do it? Do scholars have some responsibility to keep there research relevant? If I string-replace Twitter with X, how do I make clear who the original authors are, and what my role is? How do we get PLOS to point to the new version (surely not as corrigendum)? I do not have the answers.

But I do see differences between different types of research output and that makes these question an essential part of Recognition and Rewards. If it some types of output have different rules, then we do not give all scholars the same chance of recognition. Of course, this is the current situation, and just reflects that academia still has much to do to adopt Open Science.