Beginners Guide - Software Testing

Add a Comment

Comment by Dave Whalen on September 24, 2009 at 5:46pm

Software testing is like fishing. You know there are fish in there - big ones and little ones. We need to catch all the big ones and if we have time remaining we can go after some of the smaller ones. But we're not gonna catch 'em all. The key is knowing where to find the fish! You need the right sized net.

A couple of opinions:
1. Automation is great where it makes sense.
2. Automation it's expensive not just in purchasing a good product, but also in creating and maintaining tests, and in the time it takes the team to get over the learning curve to make it efficient.
3. How and when you use automation is critical. There is not much return in the early stages of development. The real returns on investment happen in the later stages as the product stablizes.
4. Some testing should never be automated - the total costs are to high. 100% automation is rarely realistic!

Comment by Joby Joseph on September 22, 2009 at 5:16am

Thanks Mr.Rob Lambert for the valuable comments.

Comment by Rob Lambert on September 21, 2009 at 11:37am

Hi Joby,

Nice post. I've added a few comments in italics:

1. Product Knowledge
Understand the product/software by referring product manuals, FAD, brows hers etc..and gain maximum knowledge about the product, which is inevitable for test plan creation and test case generation.

I guess this is assuming that tests will be written in advance. A good way of writing tests is often to use the software under test to learn more about the application. Competitor products are also a great way of finding out how things work and what differentiators are obvious

2.Think Differently
Think beyond the level of regular user. Make as many test scenarios as you can, put it together. Discuss it and remove the duplicated test scenarios. Let you have a quality test cases.

Good comment. Might be worth expanding what you mean by "Think beyond the level of regular user". Sometimes thinking as the regular user is more appropriate. Sometimes that is....

3.Target to break the product.
Understand that none of the humans in the world are error free, every one is tend to make error. So believe that, there will be some issues and you are going to break the product inorder to improve the quality of the product.

It's not really about us breaking the product. The product is already broken, we just find the breaks.

4.Use Automation Tools
Try to use the automation tools to execute the test scenarios you have made. So that will save your testing time when there is repeated testing with bug fixes and system up gradation.

Automation doesn't always save time. In fact, it often costs a lot more time and resource so use it wisely. Automation isn't just about automating tests. Automation can often be about machine control, state or time control, test reporting can be automated, data loads etc. Automation also isn't essential. I've seen projects try to shoehorn automation in because they feel it is something they should do. Spend the time wisely.

5. Generalized Test cases
Generalized test cases are the test cases which are written in a way that it can be reused for testing similar class products by adapting some changes in the master Test cases(Generalized Test cases). By this way, you could save lot of your time in writing test cases for the products under same category.

Absolutely crucial point. Admin tasks take up SO much time in software testing. I'm an advocate of reducing this waste at every possible point. And duplicate tests is the easiest one to elimate.