The OpenACS automated testing package has always been great as far as white box testing goes. We identify tcl procedures and run tests to determine if they return the expected results.
However, I don't think the automated test package helps us when it comes to black box testing. This is where we hire QA engineers to go thru the site looking for bugs and errors that our web application users and visitors will normally be the ones to experience.
Enter Selenium, I started dabbing into selenium early last year when it was still version 0.2. It is now version 0.6 and a lot of web development projects are taking an interest for various reasons.
Selenium is a testing framework created from a series of html pages and javascript. This means you can extract a copy of selenium , write tests , save it in a folder inside the selenium directory and put it on a web server where a person can get to it thru a url. See a demo here http://www.openqa.org/selenium/demos.html
I've put together a Wiki page with some information about selenium http://www.openacs.org/wiki/Selenium. I hope to later expand that and add a quick start tutorial on how I've used selenium to test some of the OpenACS packages
The 2 most compelling reasons, I think, on using selenium for automated tests are
1 - End User Participation. The potential to get end users involved in writing tests for a web application developed with the toolkit is high, specially with tools like the Selenium Recorder http://seleniumrecorder.mozdev.org/
2 - More Coverage. We're not only testing the individual tcl procedures and the api but the whole user experience with the web site/web application.
This leads me to my proposal to integrate selenium into our testing framework.
I know a TIP would be more appropriate but I would like to get the pulse of the community and how something like selenium can help increase the quality of the applications we deveolpe using the toolkit.
