https://www.youtube.com/watch?v=qDNPQo9UmJA
At 9:19 simplify UI development
A low trust and low coordination environment is where REST works well
At 1830 CRUD-flavored REST is the blog-post-execution-in-the-kingdom-of-nouns for distributed systems
At 1941 burdens the clients because you have decomposed rich interactions
At 20 we want our clients to be user advocates, thinking the way a user thanks.
At 23:50 The CRUD flavored REST obscures the story. This reminds me of why I build the availability model the way I did. It was totally unclear where or why the updates to item availability were being made.
At 31 chose Kafka for the command log, and Samsa? to process those commands
The command is a two-tuple of an action and a data payload
At 3251, these commands are rich semantic things. Can you submit them in domain terms
Om Next has CQRS Play concepts of query paths and transaction paths
At 30 for the event sourcing sounds very much like what Philip O'Toole did. Some slight sanity checks, then to Cosco
At 3920 by default the interface is a synchronous. You can optimistically update
At one point talks about Samza... Which is talked about here: blog-post-kafka-samza-unix