> Australia already has a similar system: if the research is funded by a national funder, then the pre-publisher version has to be deposited at the institute's

> Australia already has a similar system: if the research is funded by a national funder, then the pre-publisher version has to be deposited at the institute's research repository (basically the Word/PDF file you sent to the publisher, before they add their copyrighted formatting). It’s more or less the same in the UK and France as well. > I'd judge the success so/so: it's barely enforced, librarians will remind you if you didn't upload one paper, but it's not like there's any consequences. IIRC in the UK it does not count in the national performance assessment it if is not in an open repository. > It's also about the text of the paper, not about the associated data. Yeah. Data is a big problem. There are many different types that cannot be handled just like a pdf (tons of files in the same dataset, huge archives and everything in between). It’s a huge lot of storage space, the vast majority of which will never be even accessed, never mind used. I’ve been part of several initiatives and it is not easy to set up. > And these repositories are generally built by for-profit providers and aren't exactly open: I know for a fact that my Google Scholar profile does not link to the institute's repository, papers in there are barely findable. That’s a damn shame. The British government is doing a lot of things right with their open access policy, but a thing that works quite well in France is that there is a single national repository ( https://hal.science https://hal.science : it voild be better indexed but it does show up on Google Scholar) and everyone uses it. This is the sort of things that makes sense at the national level. A myriad of small repositories on obscure university websites is not very helpful. That way it also does not depend on the (possible lack of) willingness of individual institutions to fund it and do it properly. > it also doesn't change how researchers are evaluated, so the eternal rat-race to get into Nature/Science etc. continues, with all the associated problems. If we’re really serious about open access, only articles in a public repository should count. All the legal issues there used to be were cleared up, the infrastructure is there (things like arxiv can be used as a last resort for governments or institutions that won’t get their ducks in a row). The issue is really lack of good will.