Thursday, June 22, 2017

To Read Today


  • https://aeon.co/essays/what-if-jobs-are-not-the-solution-but-the-problem
    • Whoa....wild article.  I like the aggressive style.  I sent a link to Eryn to read, partially because I was thinking about her when I read it (I should be able to coast to a retirement before I'm a "surplus human") and about how I might be biased toward a traditional job >> college, major, semi white-collar job.  Some interesting quotes in the article that get to how there's no longer enough work for everyone and that it requires a shift, including a significant moral shift:
      • "Shitty jobs for everyone won’t solve any social problems we now face."
      • "almost half of existing jobs, including those involving ‘non-routine cognitive tasks’ – you know, like thinking – are at risk of death by computerisation within 20 years"
      • "solve a fake fiscal problem and create an economic surplus where we now can measure a moral deficit."
      • "everybody has doubled down on the benefits of work just as it reaches a vanishing point. "
      • "the work ethic is a death sentence because they can’t live by it."
  • https://docs.microsoft.com/en-us/azure/architecture/patterns/cqrs
    • The Martin Fowler article: https://martinfowler.com/bliki/CQRS.html
    • Command and Query Responsibility Segregation
    • I've read a bit on this before and implemented some of it without knowing the formalities.  You're separating read and update/write into read and write stores with eventual consistency, sometimes fed by an event sourcing pattern (a queue is the real data that feeds the write that feeds the read (or multiple reads).  There's a pattern very like that on one of my projects and we can replay queues to feed the write database.
    • Biggest issue: adds complexity
    • Also see Materialized Views (which only contain data necessary to a query) - this has, I think, correlations to the use of GraphQL (when the data storage doesn't serve the presentation/query).
    • Think of a materialized view as specialized cache
  • https://twitter.com/MSFTImagine/status/877692618456657920?s=09
  • https://twitter.com/thoughtworks/status/877691364665942017?s=09
  • https://stratechery.com/2017/podcasts-analytics-and-centralization/
    • Interesting to me because Erik and I built a podcast framework with tagging for professors on our product back in 2004 based on articles about the Dave Weiner work.  It was deemed a little too avant garde of a feature for our product.  Ironic, given a few years later they asked us to mimic the Apple iTunes desktop (the store was released around 2003) and aspects of our tagging and storage for podcasts were intended to be applied to other content as well and would have been a great alternative for our customers as opposed to a desktop installable we had no hope of supporting or getting budget to support.  I've worked on a few desktop install products in my time at my current employer, and they're always high overhead because of configuration differences, although it's been minimized in the last few years by providing a shell of minimal functionality that hooks back to static content served from the internet or making the app piggyback on an existing app like Outlook or Word that handles the configuration differences (hopefully).
  • https://twitter.com/barryoreilly/status/877670228804591616?s=09
    • Just a tweet about Jez Humble, but it's one of the most important rules in software.  Automate so you can focus on the fun stuff.  Including your free time and your family.

No comments: