#ASAPbio: My concerns about _The_ Central Service.

This tweet from Nick nails the TL;DR form of this blog post. You can stop here if you only want the summary.

In case you don’t follow the ASAPbio preprint world closely, there’s been a lot of discussion around the numerous preprint servers that exist. This has raised a key question: how can we, as a scientific community, store and index these documents? The ASAPbio leadership has presented The Central Service [RFA available] as the answer. I can imagine that, as a funder, it would be attractive to be able to separate legitimate preprints (which could be cited in grants) from other documents. I have a number of major concerns related to this initiative that I will discuss here.

I would like to note that one of the most common concerns that I have seen voiced (at least on twitter) against The Central Service is the argument that we already have bioRxiv. I do not find this argument compelling. I am thrilled that bioRxiv exists, and I’m glad that we get to use it. However, bioRxiv strikes me as simplified journal infrastructure with the peer review bits removed. This is great for getting a solution up quickly, and I’m sure that the maintainers will continue to innovate. However, I think that we can do better in 21st century scholarly communications and I don’t want us locked into this system. The Central Service, at least as described in the RFA, strikes me as a central repository for bioRxivs. It seems like a technical solution to the problem of “too many bioRxivs.”

Instead, what I would have liked to see out of ASAPbio as a first step would be a set of proposed requirements as to how pre-peer-reviewed documents should be identified, licensed, and archived to be cited in grants and papers. I see the need to define the degree of permanence that is required, and the extent to which a document could be changed (if at all). This discussion would have centered primarily on defining requirements that best facilitate scientific discovery (e.g. how to best support the distributed trust network of scientific communication). The outcome would have been a specific list of basic requirements that preprint servers must meet to be cited in documents submitted to the association of funders. I haven’t seen this, though there is a lot of documentation on ASAPbio.org.

Before addressing such concerns, the organization seems to have moved on to an RFA for the Central Service. This service is supposed to aggregate legitimate preprints. However, it’s not yet clear what will qualify (e.g. Figure 1). Perhaps it was possible to reach some agreement on technical aspects, while it was much harder to reach agreement on non-technical points. From the outside looking in, it feels like ASAPbio was unable to effectively agree on many of these scholarly communications questions.

Figure 1: Technical requirements are defined, however it’s clear that the standards in what meets the criteria for a legitimate preprint have not yet been set. Text obtained from the Central Service RFA.

This process, whereby standards on preprints are set by a Governance Body after the RFA is awarded, seems to put the cart well before the horse. The impending arrival of a funder-mandated Central Service will probably dissuade others from implementing their own components of the preprint ecosystem. After all, why implement an innovative aggregator if funders standardize on a different path.

I’d strongly encourage ASAPbio members and funders to consider the extent to which:

• It is important to define the minimum requirements for a research product that can be cited in proposals and papers before building a system to index these products.
• Funders will be tempted to allow only citation of products that have reached The Central Service.
• Investment in and construction of The Central Service will dissuade others from implementing independent and innovative solutions.

In summary, I don’t want to live with only bioRxiv for the next 15 years, and I also don’t want to get locked into an in theory optional but in practice funder-mandated Central Service. From a technical point of view, The Central Service seems to push preprints to a captive portal much like an early AOL. What I’d like to see is something more like the World Wide Web. I do want to see funders specify requirements on important topics like addressing, licensing, and archiving. I don’t want funders to build (and, I expect, subsequently mandate) a central service.

The current ASAPbio RFA reads more to me like a list of what funders need to count beans than what can be done to help us tell other scientists about the beans we’ve found. I have heard the argument that this is an experiment that is worth a shot. However, I suspect that once The Central Service is established it may be difficult for funders that backed the service to decide that this has been a mistake that crimps innovation.

I’d rather see a thriving ecosystem of innovative solutions that continue to improve scientific communication while also addressing funder concerns. I hope ASAPbio will decide to start with a stronger first step and aim to set standards on what pre-peer-review research products are acceptable to be cited (regardless of a specific underlying technical solution). After this is, I would like to see funders will identify and support missing pieces of this ecosystem instead of trying to construct an ecosystem from whole cloth.