How to reconcile prepub DOIs in OPS and published DOIs in OJS?

One of my OJS journals at York would like to do pre-print publications. I’m exploring whether OPS is viable solution (we’ve discounted the Forthcoming OJS plugin and we have hesitations about continuous publication).

I see via PKP’s OPS Demo that it’s possible to mint DOIs on that platform. But I understand that best practices would advise having a separate DOIs for both the prebub and published versions. So, my question is: what would be required to get the DOI to resolve to OJS once the article is published there?

As I imagine it, I could manually create a DOI in OPS based on the pattern used for that journal in OJS. I would then manually enter that DOI when entering the article in OJS. But what is then required to make switch the resolution for the DOI from the OPS prepub to the OJS published version?



Hi @tmrozewski,

Currently, Crossref’s thinking with preprint DOIs is that a new DOI should be minted for every new version of a preprint. This runs counter to the idea of one DOI for the life of a publication once it’s published. So, preprint versioning and published versioning are already a little different.

Additionally, the preprint is considered a wholly different object from the finished version and it’s likely not best that a preprint DOI resolves to the finished article. It is likely better to return to the preprint version to include the final citation as metadata. Large publishers may, for example, request the DOI for a shared preprint on submission. And, for the purpose of citations, it’s likely best that pre-print DOIs are unique in citations so that it’s clear what version of a publication was cited.

I’ve touched base with Crossref and asked that they follow up on this question as well. I know their stance on preprint DOIs has evolved a bit in the last year. I wouldn’t, in the meantime, recommend either writing suffixes manually or resolving a preprint DOI to the finished article.

Mike Nason

1 Like

Iiiinteresting - thanks, Mike! Your response has got me to clarify my thinking a little more here: the situation is perhaps more akin to “electronic first” publication that pre-prints, in that what we put in OPS would likely be the same version as what goes in OJS ultimately.

The editors of this particular journal are hesitant to consider continuous publication because of their practice of publishing both regular and special/thematic issues throughout the year.

Hi @tmrozewski and @AhemNason,

My name is Isaac. I’m the technical support manager here at Crossref. Mike’s right about our thinking on preprints. Every new version of a preprint should be registered with a new DOI, since that version of the preprint should and will be cited differently than its precursor(s). And the preprint DOIs should not resolve to the final version of record of that article.

We do recommend that each of those versions be connected to the version of record, when it is registered, using the isPreprintOf and hasPreprint relations as documented here: Structural metadata - Crossref

And here’s an example of how that is reflected in the XML:

My best,


Thanks @IFarley ! :slight_smile:

WRT the second reply from @tmrozewski, the water may be somewhat muddied by the terminology for sure. In this case a pre-print would be synonymous with “submitted manuscript” and a post-print would be the “accepted manuscript”. That version is after peer review but before copyediting. That’s likely the “electronic first” version you’re talking about.

Accepted manuscripts are often also allowed in so-called preprint repos like Arxiv by publishers, so this use-case makes sense. But they’re often an updated version of an existing pre-print/submitted manuscript.

In a way, we’re sort of talking about when something is actually published.

In this case, things are more complicated. OJS can tell you what a DOI is going to be based on the existing patterns in the DOI settings. If something is “in press” or “forthcoming” but not yet published and you want to use OPS specifically to host accepted manuscripts, I would probably use the DOI you know the article is going to have on publication. Then, when you publish, the record should update (via OJS) with the new metadata/URL. Working backwards from the publishing workflow in OJS will be easier than trying to force OJS’ hand with what’s entered in OPS.

This is sort of an in-between. And I don’t have an elegant answer about what to do with an accepted manuscript in OJS that isn’t as inelegant. Pre-print servers aren’t run by journals as part of their workflow, so I’m not sure I’d go that route in the first place. Nor do folks often upload preprints to preprint servers on an author’s behalf. I might be more likely to find a workaround in OJS for this specific instance, unless York is thinking of running a preprint server for all York scholars and they’re self-submitting.

How’s that for conjecture?! That’s a lot of conjecture!



Mike, I think you’ve hit the nail on the head. Frankly, one of the reasons I’ve hit on OPS is that I’m looking for a good excuse to host an instance at York for all of our users. In fact, they are looking for a forthcoming/in press-type feature which is a strange anachronism when you think about (funny to think how much of ScholComms is still entrenched in 17th century printing practices).

I’m going to chat with the editors later this week and reevaluate the situation. I’ll also figure out how serious an issue this is for them and how many mountains we’re collectively willing to move to accommodate what sounds to be a couple authors who really want visibility ASAP.