In the ongoing discussion about the value of glamour journals such as Nature, Science, and Cell, I think it’s worth looking back and asking: “How did they rise to prominence?” and “Are they still serving the same role they did when they arose?” So let’s take a quick trip into the history of science communication, before the internet. Then we can ask what the internet has changed, and how we could make the best of current technology.
I belong to the last generation of scientists that experienced science before the internet. I started doing research as an undergrad in 1995. At that time, I saw a web browser for the first time, and I sent my first email. While the internet had been around for a while by 1995 , its use was still very limited, and barely anybody outside academia had ever even heard of it. All this would change over the next 4-5 years, and by 2000 the internet started to become ubiquitous.
I don’t think anybody who got into science after the year 2000 can imagine what keeping up with the literature was like before the internet. I started my postdoc in 2000. During my entire postdoc time (or since), I rarely ever went into a library. By that time, most journals had a solid online presence, including back issues. Articles that weren’t available online could be requested via inter-library loan, and they would arrive electronically. By contrast, during my PhD, I spent a lot of time at the library. I would make weekly trips to browse the latest issues of the scientific journals I was interested in. Because I was working at the interface between physics and biology, I had to visit the physics library and the biology library. For certain articles I also had to go to the chemistry library. I knew exactly which library had which journals, and which journals were available in multiple libraries. (Almost all libraries had Science, for example.) To figure out whether anybody was citing a particular paper, I had to confer with the Science Citation Index, which was a big book available at some libraries. For any article of interest, it would list by which other articles it had recently been cited. Invariably, the list of citing articles would contain articles in journals that were only available in a different library on campus, or not in any library at my university. So, after reading the Science Citation Index, I’d make the trip to a different library, or submit an inter-library loan request, or make a note for my next scheduled trip to a different library to look up a particular article. In the worst-case scenario, it could take weeks until I saw a particular article, and then I’d often find out the article wasn’t really relevant to what I was doing.
Compare this to how literature search works today. I look up an article on Google Scholar, click on “Cited by” or “Related articles,” and find relevant related articles in seconds. For every article listed, I can get at least the title and the abstract, and for the vast majority of articles I can get the full text, all within a few seconds and while sitting at home on my couch. I can similarly browse any open-access journal, everything on Pubmed Central, and any paywalled journal my university has access to, from everywhere in the world. For all intents and purposes, the second I know a particular article exists, I can look it up and read it.
What has any of this to do with glamour journals? I’m going to argue that in the time before the internet, glamour journals and other highly selective journals served a crucial role. In a world where looking up a reference can take between days and weeks, finding potentially interesting references is much less valuable than finding potentially interesting articles. Yes, you would go to the trouble of hunting down a particular article if it seemed directly relevant for your own work, but you certainly wouldn’t do so just to generally keep up with a broader field, much less all of science. Therefore, reading a journal such as Science or Nature, or even a more specialized but still fairly selective journal such as Genetics, was the only way to keep up with scientific progress. You went to the library, you took the physical copy of the journal, you browsed through it, you read the interesting articles, you learned something useful, you went home/back to your office.
By contrast, these days, with nearly any article right at our fingertips, the process of publishing articles and of selecting articles can be decoupled. For example, I don’t consistently browse through the tables of contents of Nature or Science anymore, because I now have other means of discovering interesting articles. Any interesting article in my field I’ll come across sooner or later because Google Scholar recommends it to me, or somebody tells me about it in person, or it gets cited in a paper I read or review. For generally interesting articles, say the latest findings about global warming, I’ll likely see them mentioned on Twitter or Reddit. The advantage of all these methods of finding articles over browsing through tables of contents is that I’m not tied to the venue in which the article was published. I’m just as likely to come across an interesting article published in Nature as one in PLOS ONE. And the moment I learn about an article, I can read it. But imagine a print version of Twitter, in the time before the internet. If I had received per mail, once a week, a list of potentially interesting things published in the most random venues, I would never have followed up with reading any of them. The barrier to doing so would simply have been too high.
Some people take this reasoning to the extreme and argue that since now everything is easily available online and search engines are powerful, we don’t need selective journals anymore at all. The best science will rise to the top, it will be cited, tweeted, mentioned on reddit, bloggers will write posts about it, and thus we might as well publish everything in PLOS ONE. I am not entirely convinced by this argument, for the following reason: It’s all well and good if other people cite and tweet your work, but what if they don’t? If you think you have done some really outstanding work, work that deserves more attention than your regular bread-and-butter efforts do, in a world where all science is published in PLOS ONE, what options to you have? In a world that has glamour journals, it’s of course obvious what you can do: You write a nice 3-6 page summary of your work, highlighting the most important findings and the broad relevance, you send it to one or more glamour journals, and you hope for the best. If you get through, you’ll have a much higher chance of getting your article cited, tweeted, etc., because people pay attention to the glamour journals, and they like to read short, clearly written articles that highlight key findings and broader relevance. But if there are no glamour journals, then you have no good option of indicating that in your own opionion, this article is more valuable than that article.
So let me summarize the facts: (i) In the world of the internet, it doesn’t matter where something is physically published, as long as it is easily accessible through a URL. (ii) Glamour journals have lost the original purpose of making important science easily and broadly accessible. (iii) Publication venues that highlight interesting work by commenting and/or linking to it (such as Twitter, Reddit, Nature News and Views, Google Scholar, etc.) are highly valuable. (iv) Short, clearly written articles highlighting key insights and broader relevance are appreciated and highly valuable. (v) Authors have an interest in pointing out what they think are their most important works.
These facts lead me to the following proposal: Let’s take all original science out of the glamour journals. Instead, allow authors to submit short summaries (maybe 2-3 pages) of work they have already published elsewhere, e.g. in PLOS ONE. These summaries would be reviewed editorially, and also by one or two expert scientists who’d be asked to judge whether the original article appears to be scientifically sound and noteworthy. Editors might reject a summary because it isn’t deemed sufficiently interesting or novel, and there would still be fighting and politicing about getting summaries into these glamour journals, but the system would relieve authors of several pressures: (i) Authors wouldn’t have to rewrite an article multiple times just to hope to get it published eventually in one of the selective journals. (ii) Authors could get their results out and cite them properly while still trying for that glamour slot. (iii) Since the original article of record would be in PLOS ONE or PeerJ or similar, it would become generally accepted to have even the most important work published in these journals. Nobody could look at a publication list and say “Oh, it’s just a bunch of PLOS ONE papers.” (iv) Hiring committees and granting agencies would still have the option to evaluate candidates by the number of summaries they have published in glamour journals, though I would hope they would do so to a lesser degree and pay more attention to the original articles.
 The first web browser, which started the development of the modern internet, was released in 1993. The concept of an online journal was unthinkable before the invention of the web browser.