How do popular scientific news & Google get DOI links leading to zombie pages on first release, but later link to the intended scientific paper?

Question

Looking through google news feed just now I see an item that links directly to Science with the url https://www.science.org/doi/10.1126/science.abo5095 which combines science.org with a DOI. At the moment the page is empty - it contains a header and footer of the usual Science links but no article in the middle - a zombie page.

Other popular scientific news websites link to the same empty page when announcing what at least PubMed summarizes as An extremely energetic cosmic ray observed by a surface detector array, gives the abstract and links to a DOI that then goes to the zombie page.

What's so exciting is that a single 240 exa-electron volt (240E+18 eV) charged particle (possibly a proton or light nucleus, until it hit our atmosphere) had the kinetic energy of a tennis ball!

In the past (often in phys.org news items) I've clicked on links in popular scientific news articles that go directly to doi.org's "appology page" indicating that the number has not yet been assigned or linked.

In a half-day or so, these same links start working, so I need to save them and check them later.

Question: How do popular scientific news outlets and google news get DOI and DOI-like links that lead to zombie pages upon first release, but only later actually link to the intended scientific paper?

I'm interested in both:

1. the mechanism by which a DOI is issued that doesn't work but will work in the future
2. why this still happens in 2023

---

https://www.science.org/doi/10.1126/science.abo5095

Screen Shot 2023-11-25 at 9.02.52 AM.png (UTC+08):

Answer 1

This happens when a DOI or DOI-like link is prepared ahead of time, but either isn't activated or the article it points to is under some kind of embargo. The most common public-facing variant would be when such a link is included in a press release, that then gets reported on by some news outlet before the embargo is lifted. (Indeed, such news-worthy research is the most likely to be under strict embargo in the first place.) Authors can also encounter it in proofs of articles, for example.

Crossref sought input on how to handle this already back in 2016, but not much came of it. The initial report describes the mechanisms quite well

The Problem

Crossref currently recommends that publishers register DOIs at the time, or shortly after, content is made available online. This has worked pretty well over the years but changes in scholarly communications means that the current process for registering content with Crossref is increasingly causing problems both before and after publication and doesn’t meet the needs of researchers, funders, institutions and publishers.

There are a number of reasons for content to be registered prior to online availability:

1. Publishers provide authors with proofs of articles during the editorial and production process and the manuscripts have “DOI-like strings” on them. The “DOI-like strings” will become DOIs when the content is registered with Crossref but this happens after online publication. Authors sometimes try to use the unregistered DOIs from the proofs but they don’t work so this undermines trust in using DOIs as persistent links.

2. The problem described above is exacerbated when the publisher is trying to control publicity for the publication and the press (or others) are given advanced [sic] copies of the content. To those who have advanced [sic] copies, the DOI will appear to be broken and therefore the recipients are understandably reluctant to use the DOI when referring to the content in articles, reviews, press releases, etc. that are designed to be released simultaneously with the content being made available. In addition, the DOI is included on proofs and review copies sent to authors and if they try to use the DOI as a link it won’t work and they will be reluctant to use the DOI as a persistent link to the content.

3. Accepted Manuscripts in journal publishing - in journal publishing there is a lag - sometimes lengthy - between manuscript acceptance and online availability of the article (whether it be the version of record or an early publication of the manuscript “ahead of print” or being assigned to an issue). Waiting until online availability of the article or manuscript to register the content with Crossref means that a central source for an advance record of accepted manuscripts - i.e. upcoming publications - isn’t possible. Currently, there is no formal mechanism to provide metadata for manuscripts accepted for publication before they are made available online.