TIL, 2018-04-23, AI, Postgres Vacuum, Node use cases

Musings, Front-end

• Weird caching thing with Github.
• Github uses Akamai CDN for caching and services. Also, no modifying of .htaccess.
• RequireJS: A bit older, replaced by Webpack. Not that compatible with npm.
• CommonJS: A way to include JS modules within the current scope and keeps the global scope from being polluted. This reduces the chance of naming collisions and keeps code organized.
• Passing a function with parameters through props: use <Button onClick={() => this.props.myFunction(param)} />. Reference
• On parameterized event handlers: do not curry it, some binding thing can be done. Reference

Musings, Postgres:

• Multiversion Concurrency Control (MVCC): Each SQL statement sees a snapshot of data as it was some time ago, regardless of the current state of the underlying data.
  • Contrast this to locking.
• What’s the difference between a tuple and a row in Postgres?
  • Tuple is the abstract term, row is the concrete implementation.
  • n_live_tup in the pg_stat_user_tables to figure out how many tuples you have.
• SQL VACUUM:
  • Reclaims storage occupied by dead tuples.
  • Tuples that are updated or deleted are not physically removed from the table, they remain present until a VACUUM is done. So it’s necessary to do VACUUM periodically, especially on frequently-updated tables.

Musings, Ruby, Node

• Database Outage on 2016/11/28 when project_authorizations had too much bloat
  • Cool article on what they did to solve a problem with the site going down.
  • Killing queries.
  • Use htop to figure out memory usage.
  • Check dead tuples for the tables.
  • VACUUM FULL on offending table.
  • pg_repack lets you remove bloat from tables and indexes, and optionally restore the physical order of clustered indexes.
• I/O bound: The time it takes to complete a computation is determined by waiting for I/O operations to be completed, opposite of CPU bound (where computations are the ones that take longer).
• Artificial Intelligence — The Revolution Hasn’t Happened Yet
  • Distraction: the idea of having an intelligence that rivals our own.
  • Ex: the medical system that can misdiagnose unborn babies for having down syndrome, due to some error in white noise.
  • Provenance: where did data arise, what inferences were drawn from the data, and how relevant are those inferences to the present situation?
  • There is a need to develop AI systems for the medical, commerce, transportation, and education domains.
  • We are now building societal-scale decision-making systems that involve machines, humans, and the environment.
  • ML in the 90s: fraud detection, logistics-chain prediction, recommendation systems. Data science: combining ML with database experts to build scalable/robust ML systems.
  • Trend: intelligence augmentation (computation/data to augment human intelligence and creativity).
  • Intelligent infrastructure: computation + data + physical entities to make human environments more supportive, interesting, and safe.
  • Ex: data flows/data analysis flows to aid human diagnoses and providing care.
  • The scope of AI is not about science-fiction dreams or nightmares, but the need for humans to understand/shape technology as it becomes more present in our lives.