Research of potential web frameworks for use in DoIT

web_framework_research.md

Obviously most of the content on detroitmi.gov will be supported by a CMS. However, there will still be instances where we need to use a web framework to support functionality that the CMS cannot achieve. Because of this I have done some research into what frameworks would be suitable for our purposes. (Note that this is not intended to cover all commonly-used web frameworks - I have intentionally ignored some [web frameworks built in Java or C/C++, for instance] because they tend to present long development cycles that I do not feel are adequate for our purposes. Also, I am leaning strongly towards frameworks that are frequently used / well supported as well as written in a programming language that is itself frequently used, with the intention that it will be easier, when needed, for other developers to work with the frameworks.)

Some of the criteria I have for web frameworks include the following:

• should be highly productive (i.e., able to let developers achieve high-quality results in minimal time)
• should be reasonably accessible (since we have only 1 back-end developer, I have put a high priority on using frameworks that are written in well-known and readable languages)
• should use languages that are not only common, but likely to stay common going forward
• should be reasonably performant / scalable

Django (built on Python)

• pros
  • high-productivity
  • good performance / scalability
  • explicit, straightforward syntax (less 'magic' behavior than rails)
  • intended for and widely-used in publishing (e.g., used by Washington Post, National Geographic, Pinterest)
  • widely used in scientific research / engineering / data analytics (used by NASA)
  • python is especially common (more so than ruby)
  • python can easily access ArcGIS via ArcPy
• cons
  • syntax is less elegant and concise than ruby

Ruby on Rails

• pros
  • high-productivity
  • broad user base (wide variety of supported functionality)
  • concise, elegant syntax
• cons
  • sometimes confusing 'magic' behavior
  • poor performance

Grails (build on groovy)

• pros

  • high performance / scalability (runs on java VM)
  • high-productivity

• cons

  • built on relatively rarely-known language (groovy)

Laravel (built on PHP)

• pros
  • widely-known / widely supported
  • good performance
  • PHP popularity
• cons
  • PHP arguably declining in popularity and difficult to maintain