GA WDI - Week 2 - Day 4 (Thursday) Homework

gistfile1.md

DHH Offended By Proliferation of RSpec

Reference:

• Source: http://www.rubyinside.com/dhh-offended-by-rspec-debate-4610.html
• DHH: David Heinemeierhansson [http://david.heinemeierhansson.com]
• Test::Unit: http://www.ruby-doc.org/stdlib-2.0.0/libdoc/test/unit/rdoc/Test/Unit.html
• RSpec: https://github.com/rspec/rspec

Issue(s):

• RSpec is overcomplicated - offends him aesthetically
• There is no discernable benefit for its added complexity
• DHH respects experimentation, but is sad by the proliferation of RSpec

Public opinion:

• Gist from DHH: https://gist.github.com/dhh/893027
  • Is DHH being disingenuous in his example code to support Test::Unit over RSpec?
• Test::Unit
  • Feels right - simplicity has a huge aesthetic value
  • Is easier to get started with...
  • Code example...
    • if assert respond_with(:success) fails, the two assertions that come after it will not run.
    • this is not the case with the RSpec example. So Rspec can seem more robust...
• RSpec:
  • Is not about testing, it is about documentation
  • Sometimes it's easier just to have it all in one spec.
    • Of course personal taste plays into this.
• The examples look relatively similar - it boils down to do you prefer:
  • [RSpec] 'assert something' ; or
  • [Test::Unit] 'object.should be_true'
• The closer [the code] is to the way I'd describe the behaviour in English, the less likely I am to miss a mistake.
  • [RSpec] something.should equal(:something_else) sounds a hell of a lot more natural
  • [Test::Unit] assert_equals something, :something_else sounds very robotic and requires me to translate from computer language to human language, however briefly.

My viewpoint: RSpec seems like the better choice...

• I agree that it is better to have your code reflect how you would describe it in English
• RSpec seems to provide a more natural form of in-built documentation
• RSpec test specs seem more robust, in that they run to completion, not failing to run if a previous assertion [Test::Unit] fails
• Understanding the code should take priority over aesthetics, especially in a shared-code development environmnet
• The initial learning curve is advocated as being harder for RSpec, but that shouldn't be a reason to adopt Test::Unit
• Both approaches can yield equivalent results, if you understand well how to use both systems. After that it come down to how useful it is and the concensus/preferences for the group of developers handling the code as a whole.
• There should be standardization on one methodology to keep things consistent.