I think the saying goes: fool me once, shame on you; fool me for like 7 or 8 years in a row, shame on me. At this point, if MongoDB could turn on a dime from b

I think the saying goes: fool me once, shame on you; fool me for like 7 or 8 years in a row, shame on me. At this point, if MongoDB could turn on a dime from being an unreliable system to a robust one, that would be the most remarkable part of this story, because I've never seen that happen with any engineering organization. It's rare that engineering departments can significantly improve their quality culture, and it's never fast. Microsoft Windows went from crashing (for me) >=3 times a day to crashing almost never, and it took them over 10 years -- and replacing their kernel twice. I've seen nothing from MongoDB to make me believe they even realize the problem, much less are on their way to solving it. In the lifetime of a database, 2 years is nothing. And it's always easier for quality to slip than to improve. If they turned it all around since I last touched MongoDB, good for them, but they've burned through all their trust, and it's going to take a lot longer than that to earn it back.