Some researchers make hard assumptions about the failure models (only
fail-silent) and on the desired schedule (only static)

Able to
generate fault-tolerant schedules
automatically

For
instance: [Kulkarni 00], [Bertossi 99], [Caccamo 98], [Ramamritham
95], ...
Other researchers have much less restrictive assumption

But their
solutions are
hand made

Specific
communication protocols, voting mechanisms,...