We should do more benchmarking and profiling. Where are currently bottlenecks? What's the throughput of these changes? How does it compare with other similar systems (Meteor)?
web-mongo copied to clipboard
mitar
