I’m tentatively excited about keybase’s new filesystem, but I wonder if some of that excitement is simply because their directory structure — where I have a /keybase/public/<identifier> hierachy that can be mounted by anyone, and a /keybase/private/<me> folder that is synced only between machines I attest as controlling — maps so well to the structure I’ve been trying to use in my own home directories for, gosh, over a decade.
The top-level directory in my ~danny/ has a Private and a Public folder. The Private directory is encrypted, and is linked into by a menagerie of symlinks whenever I find something that I wouldn’t want the world to see, from configurations to tax documents. The Public folder, in theory, contains everything I wouldn’t care the world seeing. My ideal was that I’d just share ~/Public on a webserver, and I’d try to err on the Public side. In practice, I’ve never actually been brave enough to open up all of ~/Public. Too much private stuff gets emitted, even accidentally. As I was writing this, for instance, I realized that I had half-written a script that could be used to derive a relatively important password, and it was still slumped around in Public (I’ve always tried to keep all my ongoing code repositories on the Public side). Just the idea of auditing the vast stash that has mounted up in there has lead to me growing ever more cautious.
I wish there was some middle ground between those two folders. But there isn’t, and that’s the world we live in. Unless I should mkdir ~/Obscurity one of these days.