If supplying eprints to requesters could be delegated to 3rd parties like Repository Managers to perform automatically, then they would become violations of copyright contracts.
What makes the eprint-request Button legal is the fact that it is the author who decides, in each individual instance, whether or not to comply with an individual eprint request for his own work; it does not happen automatically.
Think about it: If it were just the fact of requesters having to do two keystrokes for access instead of just one (OA), then the compliance keystroke might as well have been done by software rather than the Repository Manager! And that would certainly not be compliance with a publisher OA embargo. “Almost OA” would just become 2-stroke OA.
No. What makes the eprint-request Button both legal and subversive is that it is not 3rd-party piracy (by either a Repository Manager or an automatic computer programme) but 1st-party provision of individual copies, to individual requesters, for research purposes, by the author, in each individual instance: the latter alone continues the long accepted tradition of reprint-provision by scholars and scientists to their own work.
If reprint-request cards had been mailed instead to 3rd-parties who simply photocopied anyone’s articles and mailed them to requesters (with or without a fee) the practice would have been attacked in the courts by publishers as piracy long ago.
The best way to undermine the Button as a remedy against publisher OA mandates, and to empower the publishing lobby to block it, would be to conflate it with 2-stroke 3rd-party OA!
That practice should never be recommended.
Rather, make crystal clear the fundamental difference between 1st-party give-away and 3rd-party rip-off.
[Parenthetically: Of course it is true that all these legal and technical distinctions are trivial nonsense! It is an ineluctable fact that the online PostGutenberg medium has made technically and economically possible and easily feasible what was technically and economically impossible in the Gutenberg medium: to make all refereed research articles — each, without exception, an author give-away, written purely for research impact rather than royalty income — immediately accessible to all would-be users, not just to subscribers: OA. That outcome is both optimal and inevitable for research; researchers; their institutions; their funders; the R&D industry; students; teachers; journalists; the developing world; access-denied scholars and scientists; the general public; research uptake, productivity, impact and progress; and the tax-payers who fund the research. The only parties with whose interests that optimal outcome is in conflict are the refereed-research publishers who had been providing an essential service to research in the Gutenberg era. It is that publishing “tail” that is now trying to wag the research “dog,” to deter and delay what is optimal and inevitable for research for as long as possible, by invoking Gutenberg-era pseudo-legal pseudo-technicalities to try to embargo OA, by holding it hostage to their accustomed revenue streams and modus operandi. OA mandates, the immediate-deposit clause, and the eprint-request Button are the research community’s means of mooting these delay tactics and hastening the natural evolution to the optimal and inevitable outcome in the PostGutenberg era.]
Sale, A., Couture, M., Rodrigues, E., Carr, L. and Harnad, S. (2012) Open Access Mandates and the “Fair Dealing” Button. In: Dynamic Fair Dealing: Creating Canadian Culture Online (Rosemary J. Coombe & Darren Wershler, Eds.)