cmdrkeene

Here's a quick ruby implementation:

https://github.com/cmdrkeene/bitmap-counter

GroupMe 2.0 (released today) uses either SMS or Push/TCP in the iPhone and Android apps.

To your issues:

1) Yeah, organization is crazy. I think the method comparison is a bit unfair though. In theory, each step should be easily identifiable (like a method) but this is rarely the case given that most steps have complex regex matching (sometimes for grammatical niceness - e.g. (a vs. an), sometimes for other crazy town reasons). This makes not only automated indexing difficult (RubyMine tries...) but trying to keep hundreds of steps in your head is pretty hard. Especially one's you probably didn't write.

The cool thing about steak is that helpers and "reusable steps" are actually just methods! Hello ctags :D

b) Our case was:

  And the user "212-555-1234" should receive a text with:
    | Some "double quoted" and 'single quoted' text |

  text.should == "..." (you have to do this anyway in the step_definition)

The trouble is you have a url:

  http://coolsite.org/something/abcde22424aw3324234

I click on the link "Link Title or URL". Since the token is random every time (unless you stub it, which is kinda out of the spirit of acceptance testing, but that's another debate) it's hard to say "Click on the link with the big ass token in it".

You can definitely write a step like:

  And I visit the account confirmation link

d) Similar to the above, the catch all step can work in cucumber. Though, in a few scenarios, it's desirable to run only specific jobs or more often in our case, to check if jobs are scheduled in a specific way. It avoids the weird scenario of writing a step definition like "And I run only the blah jobs" or "And a blah job should be scheduled with ...". Ideally the job system would be decoupled entirely from the acceptance test, but we can't all do cool stuff like this all the time: http://corner.squareup.com/2010/08/cucumber-and-resque.html

Steak doesn't want to be cucumber at all. It's plain ruby. The "DSL" aspects of the library are purely for legibility and structure. The keywords feature, background, scenario perform the exact same function as the RSpec keywords describe, before, and it respectively.

All steak does is provide a structure for acceptance tests and loop in a web automation driver (capybara).

Far less magical than parsing a the magic language of "Gherkin"

To your points:

1) Agreed. It's simple. But, it does some nice things like 1) hook up capybara 2) create a /spec/acceptance folder and suite hook and 3) add some "behavior-oriented" keywords like feature, background and scenario. Of course, these are just dumb blocks and yes you could use the provided RSpec DSL (describe, it, before, etc) but it helps (if only mildly) to indicate that this is an acceptance test and to maybe put the programmer more in that mindset. This isn't groundbreaking, it's just a nice convention in my opinion.

2) This is a fair criticism. It does put the programmer more in a "code-centric" mindset than a "behavior-centric" one.

To address this I usually write out the feature in comments exactly as I would in cucumber. This might sound like an argument against steak, but it's not. The most useful thing about cucumber to me is the pattern of thinking: "Given this, When I do something, Then I should see that". The trouble is in the details: the maintenance, the awkward step gymnastics, etc. If you write out a pseudo-feature in comments then under each line of comments write the code, it totally remove the layer of feature vs. step definition indirection that plagues large cucumber projects.

3) Totally agreed. Though, very few projects I've used cucumber on ever let customers see feature files, let alone write them. Steak is for programmers, not customers.

There is no issue ramping up because it's just rspec and capybara. This is exactly what's under the sheets of most cucumber (for rails) installs. Anyone who implements cucumber step definitions should be well versed.