Why do researchers petition for Open Access (OA) yet fail to provide it for themselves, by self-archiving? An employer/funder self-archiving mandate is what is missing to resolve this koan. What needs to be mandated is only the keystrokes for depositing the final draft plus the OAI metadata of the article in the author's Institutional Repository (IR) immediately upon acceptance for publication, along with the strong encouragement to set access-privileges as OA. Access to over 90% of these articles can already be set as OA with the blessing of their publishers. The rest can be set to IR-internal access for the time being, but their metadata will still be as visible to all searchers and surfers webwide as those of the 90% that are already OA, allowing would-be users to email the author to request an eprint. Emailing eprints can bridge the gap until either the remaining 10% of journals give self-archiving their blessing or the author tires of doing the superfluous keystrokes to email the eprints and simply does the last keystroke to set access at OA. Either way, mediated OA will already be providing effective 100% OA as of the implementation of the keystroke-policy. Such an immediate-deposit ("keystroke") policy -- leaving no loopholes for any exceptions or delays -- is what Research Councils UK (RCUK) needs to mandate. The rest of the planet will follow suit. And Nature will take care of the rest.