[erlang-questions] obviously no bugs? (Re: Alternative supervision approaches)

[Raoul Duke]

> Stop dreaming and start building real things, I say.

i don't disagree. :-) i don't think i said, i wasn't trying to say
that monitoring and restarting is something we can do away with.

i hold out weird hope for more formal methods someday automagically
helping out with the failures that we actually *could* study 'on
paper' and thus avoid. yes there will always the tripping over the
power cord type stuff even though. still even there i'd love it if my
model checking could tell me, "dood, if the power goes out at this
spot of the code, then your db will be totally horked and your
recovery will be a nightmare" etc.)

e.g.
this:
http://www.hpl.hp.com/techreports/2011/HPL-2011-44.html
is not at the formal methods level but it seems like a nice attempt to
incrementally do more study and nailing down of such issues above and
beyond the run-of-the-mill restart-and-pray stuff.