While we talk about the future of science publishing, we need to also have a conversation about how to fund it. The problems with the subscription model is obvious and needs no elaboration. But it's clear that the APC/author pays model doesn't really work either.

Back at the dawn of the #openaccess era, when @BioMedCentral and @PLOS adopted APCs, we knew it was imperfect, but it was the only viable way to cover costs that didn't require locking papers behind paywalls.

It was our expectation that this was a transitional state - that the funders who ultimately provide the money for science publishing would realize that it doesn't make sense to fund research infrastructure like publishing with transaction fees of any kind.

Funding journals via subscriptions requires paywalls and excludes readers. Funding journals via APCs encourages high volumes, and erects barriers at the point of publication (although these are generally mitigated, albeit imperfectly, with discounts and waivers).

And you can see an additional problem with APCs in the new initiatives @eLife is taking. If our primary job is to review papers, and not publish them, then it makes no sense for authors to pay only if we accept their paper.

In principle we could move to charging for review, and there are some good arguments for doing this. But I don't think this is optimal either. To me it ultimately makes far more sense to dispense with the conceptualization that science publishing is a bottom line business.

The reality is that science communication is research infrastructure - more like a genome database that a reagent supplier - and it would be better for science in a million ways if the funders who sponsor research recognized this and develop a way to directly subsidize publishing

For example, right now it's free for authors to post papers on @biorxivpreprint because @cziscience and @CSHL subsidize the process. This is how it should be! It would be counterproductive if people had to pay to post preprints - even if the fee were nominal.

The same thing should be true of peer review. This is critical research infrastructure, and for far too long we have let the goals of the marketplace dictate how peer review should be organized and funded. And the net result is a system that costs us $10b/year and doesn't work.

I'm not sure exactly how we should structure a subsidy the system so that it's fair, efficient and encourages good service and innovation. But I know that neither subscriptions nor APCs do that, and it's time for us to figure out a better way.