[rspec-users] I lost the RSpec fight

Jay Levitt lists-rspec at shopwatch.org
Sun Sep 16 10:44:19 EDT 2007


On 9/16/2007 10:18 AM, Kevin Williams wrote:
> By all means, they should not ever try anything new. The people of
> Earth should not have ever adopted the use of the electric light bulb
> or the radio or the automobile or the airplane or the microwave or the
> telephone or ... the Internet. No, don't adopt anything new, just
> stick to the old ways of doing things. It's impossible for anyone to
> improve on anything, right? Bah! Humbug!

Obviously, I'm with you on that - but I wanted to make the point that, 
to make inroads in the greater community, "change is good" is not a good 
enough mantra!

One of the problems is that, like templates (and yes, they're still on 
erb, if you mean as opposed to HAML), you can't really use RSpec for 
part of a project.  There are no technical barriers; rspec and 
Test::Unit sit side-by-side quite nicely.  But procedurally, 
organizationally, it's a pain.  And I don't think there's a way to ease 
that pain.

A few more impressions from that meeting:

* RSpec might be nice, but Test::Unit is the "least common denominator". 
  We all know it, it works with everything, it's there out of the box. 
RSpec doesn't have enough compelling reasons to change to it.  (Maybe an 
updated feature comparison would be good here, once 1.1 launches 
w/StoryRunner and perhaps these HTML Formatter changes.)

* I mentioned how Test::Unit development had stagnated, and they took a 
different view of it:  Test::Unit is "complete".  It's done, it works, 
it provides a base level of functionality that doesn't need any more 
updating.

Jay



More information about the rspec-users mailing list