5 ms·
Postgres wasn't even on my radar until the 9.1 release (2011) rounded out their replication story. MySQL replication had been baking for a decade at that point,
by tomnipotent 3y ago
Postgres wasn't even on my radar until the 9.1 release (2011) rounded out their replication story. MySQL replication had been baking for a decade at that point, despite its other glaring flaws.
- SOLAR_FIELDS 3y agoYeah, in my personal anecdotal experience Postgres wasn’t really known as the “good enough for anything, rarely the best at anything but pretty good at it” workhorse until the mid 2010’s. It is great though, amazing piece of technology. Not 100% sure you need some specialized data store for your use case? (Most people aren’t). Pick Postgres. You can always switch to SpecializedDB later when pg falls over. I cannot undersell enough how hard it is to be at being 90% good at everything. Like, you can pick Postgres for these technologies and not get fired for it at first glance and it will carry you to medium-large scale - time series - GIS - Graph - CRUD/ Jamstack apps (eg Supabase) - Queues - and on and on and on If it even tangentially involves data you can trust that Postgres has some solution that will mostly work with some edge cases unless you are at disgusting levels of scale. Love it. A true jack of all trades