2 items
-
#1
Distributed Erlang
Erlang’s actor model, OTP, and distribution work together beautifully to build fault-tolerant systems—but there are gotchas.
From process mailboxes and supervision trees to clustering challenges, network partitions, and the single mailbox bottleneck.
This post explores the many strengths, and subtle pitfalls, of building distributed systems with Erlang.
-
#2
You have built an Erlang
You wanted a simple service notification system, so you added HTTP callbacks, then queues, then retries, then supervision…
Congratulations! You’ve built an ad-hoc, informally-specified, bug-ridden slow implementation of half of Erlang.
A satirical journey through accidentally reimplementing the actor model as seen on Hacker News.