If I coud say anything, AutoIT is very good for simulating user interaction – quite simple and rapid. This kind of tests are expensive to maintain though… There are lots of external issues that can result in an AutoIT test fails (e.g. another process or a pop-up that steals the focus, when the script is supposed to click something).

IMHO it’s better to write as much of simple unit tests as possible (test the code) and use AutoIT for cases when user interaction is the aspect to be tested.

But anyway, AutoIT has it – it’s very powerful. You can do a lot with it!

Jarek

This article is great, however, it’s rather simplistic on test automation. One of the big problems with test automation is detecting or validating the state of the GUI. It is much easier to just drive the GUI without checking its state.

Being a free tool, AutoIt is more limited in its ability to detect the state of GUI components. Commercial tools offer you additional ways to do the detection, and may be more robust in the detection process.

And last thing to mention, GUI component (IDs or names) can change in every build/release and thus can break automation because you’ll have to update the IDs/names of the controls in the automation code.

I have something to suggest; I’ve been using a similar approach, but with a commercial product, albeit a very inexpensive one at that…

Why Homebrew + Tool in development?

First, regarding the HomeBrew solution, well this article says it all. Much like your approach, where you aim to cover more in less time. HomeBrew

However, what if you would like to do the same for your company/testing but you’re not a programmer (like myself), or never, ever had anything to do with programming. Is this road totally off limits?! Well, not completely. I came across an automation tool, Macro Scheduler that works extremely well because it follows your idea and comes complete with the pre-built functions (like the ones you’ve created) and allows the tester create scripts that wait for windows, and log error messages easily.

I think that a powerful and quick scripting language is best, and it seems impossible to achieve that in short time. But a good High level language such as Perl, python, Smalltalk can be leveraged to achieve great results in the right hands. But what about making it work, Debugging, etc.?

These aren’t easy things for some people to learn. I was totally lost in college when it came to programming. C++ was an F-0 for me. So what does somebody like me do?

Go into film making?

Uh, maybe…

Then I came across Macro Scheduler. I love it because it allowed me to, as I mentioned already, create scripts to do the same while I’m just a mere mortal.

-I have to admit, that the code above is overwhelming for a guy like me, but I still want and need a better way to do my job. So what’s the advantage that you get with Macro Scheduler.

Well, its THE SUPPORT; the number one reason someone like me might need a commercial tool, and this justifies the cost. Actually, I think the tool is free for all intent and purposes, and the amount paid is for the support received. You get many examples and a great support community, along with a tremendous support effort from the company itself (very high quality indeed). All this make it possible for others (like myself) to, for very little money mind you, do the testing you’ve suggest quicker and easier.