Home > Archive > Software Testing > April 2006 > [ANN] browserUnit: automating internet explorer to test web pages
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
[ANN] browserUnit: automating internet explorer to test web pages
|
|
| Achim Domma 2006-04-12, 7:08 pm |
| browserUnit
-----------
browserUnit is a C# library to automate the internet explorer and to
simulate user sessions. It can be combined with NUnit to easily set up a
test suite for web applications.
The tool is mostly inspired by the Ruby library Watir, which works very
well but does not fit my .Net environment. Due to the lack of a better
free tool, I started developing one by my own. The project can be found at
http://browserunit.sourceforge.net/
Any feedback would be very welcome!
regards,
Achim
| |
|
| I forwarded this to testfirstuserinterfaces@yahoogroups.com, and
agile-testing@yahoogroups.com .
> The tool is mostly inspired by the Ruby library Watir, which works very
> well but does not fit my .Net environment.
Now the "problem" here (if I might guess at a problem without trying either
tool) is that both snuggle up with Internet Explorer and neglect all the
other browsers. I'm aware for most situations any tests are better than
none, and a good test rig will indirectly enforce compatibility with other
browsers.
But with AJAX turning our thin clients back into thick ones, I suspect the
next level for all these testers should be integrated compliance
verification. I often plug Tidy into such test rigs, to at least enforce
standardized XHTML outputs.
It's just a thought; any new *unit system should focus on its early adopters
first, and my ramblings last!
--
Phlip
[url]http://www.greencheese.org/Z Land[/url] <-- NOT a blog!!!
| |
| Achim Domma 2006-04-13, 7:08 pm |
| Phlip wrote:
> Now the "problem" here (if I might guess at a problem without trying either
> tool) is that both snuggle up with Internet Explorer and neglect all the
> other browsers.
That's true, but I try to keep the interface browser neutral. As far as
I know it's possible to automate Firefox via a Telnet connection. Might
be possible to integrate such functionality, but I don't have the time
to do it.
It's not such a problem for my tests, because I'm testing bussiness
logic on the server in the first place.
> It's just a thought; any new *unit system should focus on its early adopters
> first, and my ramblings last!
And it should focus on the main problems to test. I'm focused on testing
the workflow or user stories of my application. I'm don't care too much
about layout differences. That might be important for some users, but
it's not my first target.
regards,
Achim
| |
|
| Achim Domma wrote:
> It's not such a problem for my tests, because I'm testing bussiness
> logic on the server in the first place.
Tests should be as close to the testee as possible.
Logic Layer ->
Representation Layer ->
GUI Layer ->
Web Browser
The "GUI Layer" is your server-side code that drives some system to generate
HTML.
The "Presentation Layer" is a layer devoid of GUI Toolkit identifiers. It
exposes the same interface to programmers as the GUI does to users. If the
user could click on a list, then the Presentation Layer has a method that
returns the list, as an array of strings or something, and a method that
takes one element of the list and processes it.
The "Logic Layer" is the database, business rules, and such. It might not
have any list in the same format as the user sees. The Presentation Layer
translates from logically convenient formats into user-convenient formats.
You need tests on the Presentation Layer, and you need to decouple it from
your GUI Layer. That should be thin.
Testing a Logic Layer or Presentation Layer thru a GUI is an industry
anti-pattern. Yes, it will often work very well. It adds fragility and
coupling to a project.
--
Phlip
[url]http://www.greencheese.org/ZLand[/url] <-- NOT a blog!!!
|
|
|
|
|