In his talk on screen recording APIs, Jason Horn mentioned he's an avid fan of Sauce Labs and their hosted Selenium solutions. There are a host of features and pricing starts from free (Linux and FireFox, limited to 500 min/month) and goes up to monthly and enterprise pricing. You can try some of the premium services with a free trial if you want to check it out. While I haven't personally used the product, I really like the idea and I'll have to check them out on an upcoming project.

For those of you who do web testing, Support Details is a handy little utility for when you report issues. It automatically pulls down information related to the browser you're using, and it allows you to either download that info (csv, pdf) or email it to someone who can look into the issue. Since I have several computers and each has a couple of browsers (all constantly doing their own automatic updates) this is really nice for when I submit something that I suspect might be browser specific.

In the good old days of the Web 1.0 world, when something didn't go well you'd get the basic 404. On a fancy site, that 404 would perhaps be branded. In today's Web 2.0 world, 404 pages are often smart. They try to figure out what you were doing and suggest places where you might go next. It's a cool feature, and a great way to take what might have been a poor user experience and turn it into something positive.

However, it doesn't always work the way we intended. When sites suggest things to you, they can make mistakes. Therefore, I have a heuristic that I follow which says always go where the software suggests I go. Here's an example...

Last week while playing around with the new mint data website, I was greeted with the "City not found" page several times.

A feature of this page is that it suggests a page for your to visit next. Often, it's a state. However at point point I was able to get the site to suggest I view spending for the entire USA. Using my heuristic, I followed the link and was greeted with an excellent java exception.

Rewarded with this exception, I was able to determine that the site is written using Java (sprint and hibernate), runs on Apache (and the version of Apache). I also have suspicion of what to do next. After seeing this issue I started trying custom URLs and was able to get several different exceptions. Some of which gave me additional insight into the site structure and possible test ideas.

Total Validator provides the following main features:

• A parser that validates the basic construction of your pages


• True HTML validation against the W3C Markup Specifications or ISO/IEC definition using the published DTDs and standards: (2.0, 3.2, 4.0, 4.01, 5, ISO/IEC, XHTML 1.0, 1.1 and 5, XHTML Basic 1.0, 1.1, (X)HTML+RDFa, XHTML-Print)


• An accessibility validator that validates against the W3C Web Accessibility Guidelines (1.0 and 2.0) and US Section 508 Standard


• A broken links validator that checks each page for broken links


• A spelling validator that spell checks the content of your pages (English, French, Italian, Spanish, German)


• Snapshots (screenshots) of your pages in different browsers, on different platforms, at different resolutions


• A desktop tool so you can validate pages before you publish, and pages behind firewalls

Here are results generated after our home page validation.

The picture below is a screenshot of a web-site opened in 4 different browsers. There is nothing unusual in that - just a regular web browser compatibility testing practice.


What is unusual - none of those browsers was actually installed on test machine.

This is a free service by Spoon: you can invoke multiple copies of different versions of different browsers without having them on your machine.

What's currently provided

• Microsoft Internet Explorer (versions 6, 7, 8, 9 Preview)


• Mozilla Firefox (versions 2, 3, 3.5, 3.6, 4 Beta)


• Google Chrome (versions 5, 6 Beta)


• Apple Safari (versions 3, 4, 5)


• Opera (versions 9, 10)