I like to see this on a spectrum on 'reliable computation' and 'high availability' (and interruptible computing) with continuations/CallCC, checkpoint/restore (

I like to see this on a spectrum on 'reliable computation' and 'high availability' (and interruptible computing) with continuations/CallCC, checkpoint/restore (CRIU, DMTCP), VM snapshots or (live-)migration, to double/triple hardware, and even dual implementation, but also Kafka. I find it a bit disheartening that these features are not more pervasive on language, runtimes, libraries and OSes, so very often the actual solutions are either very kludgy or very high maintenance, whereas the primitives are really interesting for many other endeavours.