no fluff: day 2

well, on to the next conference!  after oscon, i traveled back home, saw a movie with my wife, woke up and headed out to the no fluff just stuff conference in phoenix this weekend.  because of oscon, i missed the first day of no fluff...

so the no fluff day 2 started for me by going to see a session on testing with selenium by neal ford.  selenium is a web testing framework mainly designed for user testing (i.e., user acceptance, use cases, etc.) -- *not* for unit testing.  selenium is a completely javascript application, which means it executes on the client and no interaction with server process.  it implements iframes, etc. to display a testing control panel and viewing your application, etc.

selenium was originated like some other things lately (rails, etc.) where it was extracted from an application use -- meaning it wasn't developed expressly for the purpose of being developed.  the name came from an interesting jab at mercury (selenium is an element known to help mercury poisoning).

i was very impressed with what i saw -- very cool stuff and well thought out testing.  the downsides are client-based only, but that is the purpose of the tool.  because it does things on the client, there is no auto log shipping to any shared file.  however, there is an option to provide a url to post all completed results which will help for that logging. 

i would definitely look at this platform for future projects.  there are some benefits that visual studio team test has, but selenium also is superior in other ways such as ajax testing.  i'd love to see some selenium test on atlas as it is able to handle async postback testing, etc.  neal did a demo using google maps that was able to read the elements, etc. very easily.  there is also a firefox plugin to record a selenium test and save as the normal html test cases or as java, python, perl, c#, or ruby.  very cool.