Scales to very high numbers of connected clients, and is faster on the low end and faster on the high end. Beating both the default per-thread and the Enterprise Edition thread_pool connection handler. Enjoy!
Creator of Apache CouchDB here. Fireproof is a very clever system for building secure, collaborative apps. Your apps can work offline and when connected automatically sync changes to local storage and display them immediately to the UI. It has much of that same CouchDB magic with almost no complexity or headache on the backend.
My understanding is it’s from car racing. Putting sandbags in during qualifying to make min weight. Then removed during the race to make the car lighter and faster.
Anecdote: about 2-3 years ago I was using about 10g/day kratom and 1gm/day caffeine to deal with chronic fatigue and muscle pain.
I was dependent on both for about 12 months just to function (without them I had a hard time getting out of bed or anything done). But the combo helped to finally get regular exercise, which helped with sleep, which helped me to heal.
The caffeine was _much_ harder to kick than the kratom. A full week of headaches and fatigue due to caffeine withdrawal.
When people talk of kratom being bad because it produces dependency, why don’t they talk of caffeine in the same way? What real negative effects does the kratom have if used long term? My experience says there is none. And all this handwringing is simply due to lack of familiarity.
It _is_ append only. But doesn’t use flat files. The main storage files have 2 different immutable btrees to find documents (by id and by sequence num). And it does have mvcc for both storage and secondary indexes. Internally it has transactions but they aren’t exposed to the user, as multi document transactions don’t make sense for the replicated document model.
Really? A spiders reaction to caffeine is your reason? I just quit caffeine. I'm ready for some conformation bias to justify my decision, but this is pretty weak.