Solving big async problems by thinking differently about events.
Variables to tracking the state of async stuff
1994 - Design pattern book
Both Iterator and Observer: Producer is sending data to the consumer one piece of data at a time.
In Observer, data producer has no way of telling the consumer that sending data is done.
What is difference between array and event? Both are collections, this mindset make async programming easier.
Observable - collection, over time.
All tests in Jest run in parallel, it can’t interfere.
Automocking - auto isolation.
Clear feedback about failing messages.
Understand module system and their dependencies.
Fast iteration, instant feedback.
Single node key-value store.
Keys are sorted - possible "ray scans”
Open source by Facebook, forked by LevelDB
Especially good for map (spatial) data.
Spatial data are usually optimized for reads, cause import is one big import of all data.
> We hate routers, load balancers are stupid :D
Challenges about deploying hardware around the world.
Dave left netflix because everything was moving to AWS and it wasn't interesting for him.
Specialized to do one thing well - deliver video.
Control the path between Netflix and end users.
When entering new markets - you need to know the intersection of network and even the facilities.
It needs to be easily to deploy by contractors.
= micro cluster
uggh, … to hardcore for me
Fighting spam at Facebook for 4 years
Facebook platform - how data gets in and out of Facebook Open Graph.
* API + Application Model
* Social Plugin
It can reach my friends.
Solution: Review all the apps. Facebook was developed as open platform.
Prevent impact of false possitive.
Victim get hacked by clicking on e.g. Like button.
UI Redress attack.
Suspicious name :D
Following the mouse.