Django History

Considerations

• Web development needs to be stupidly fast
• automate repetitive tasks

• best practices

  • Obsessed with web standards and best practices
  • You have to enhance, extend, and maintain

Choices

• open source
• BSD license
• python

• pragmatism over methodology

  • Didn’t care about GoF, MVC, TDD?

• Why break the standards?

  • A lot of it is snake oil
  • “There is no silver bullet” - Fred Brooks
  • There is the right way and getting things done.

• Ramifications on questioning anything new

  • “What’s the need?”
  • No architecture astronauts
  • You risk reinventing the wheel
  • You can take things to far (Ruby for Rails)
  • Messy internals

• Web development ought to be fun

  • Rails asks why software development can’t be fun
  • APIs should match behavior, not show how they work

• Good APIs almost documents itself

  • Documentation driven development “DDD”

• Why does Django ship full stack?

  • Lack of choices at the time of inception
  • Might make separate choices again
  • There is a reason why glue code is called glue code
  • Full control over code means you get to define the API nicely

• Don’t repeat Zope’s mistake

  • Don’t have people learn Django and not Python
  • Innovation happens elsewhere
  • Always someone smarter outside your own group
  • Small pieces advance quicker

• Don’t make opinionated software

  • Ego is dangerous
  • It is not about you
  • You can’t possibly predict what everyone wants
  • There is no “one obvious way”

• Lots of choices is bad for new users

  • There are always more beginners than experts for your projects
  • Make complicated things easy

• Choices in the future

  • Long tail or huge scale
  • Learning curves
  • Django in the Enterprise
  • Does Django want to fight STRUTS

  • Avoiding inevitable backlash

    • Inviting criticism and take it well
    • Make conscious decisions