Below is an excert from another post. This spawned a question that I felt deserved it's own post.
<BLOCKQUOTE><font size="1" face="Verdana, Arial, Helvetica">quote:</font><HR>Originally posted by allogene: We do not use the best method approach here, but generally we start with one Analyst and one tester. The analyst writes the test plan and begins manual testing while the tester begins on automated scripts. Once the automated scripts for functionality are done, load scripts are created if requested (its bad but not all of our products get load tested). Generally we will add an extra Analyst to help out if need be, but usually there is only one tester per project. Very rarely do we have more than one.
My theory is that no AUT, no matter how small, should have one set of eyes looking at it. How can one person with their limited scope and single skill set possibly test an app. I would think that a minimum of two should be used for the testing, not just the planning. One person will never test the same as the other, it is kind of like finger prints. They may be close but they are always different.
My question(s) to the rest of you is what do you think?
Is one person testing OK? If not what would you consider the minimum?
Is one person planning the testing OK? If not what would you consider the minimum?
I wish things were different, believe me, but I do not make those choices. The biggest reason our system is set up this way is time constraints. We generally do not have the time to have more than one person write a test plan etc. And usually the same people work on that project for the life of the project, meaning the different versions of it. They become very familiar with it and know what to test for.
I know this is not always good because they can become to familiar with something and not test something because "it already works." But with a staff of about 12 who work on more than 40+ projects it the best we can do.
While I'd tend to say no it depends on the type of product. What if there is only 1 developer writing the product? While it is true that no 2 people test the same way, a good tester with proper skills and background will be able to apply techniques that will enable then to feel confident in the coverage the code gets, but that depends on how big the project is and how much time is available to test. Do I feel 1 tester is capable of testing the space shuttle navigation program in a month? No. Do i feel 1 tester is enough to test a website in 2 months, yes. Bottom line is that it really depends on the situation and other weighted factors.
I always make sure one tester is responsible for a set of tests, however, I also make sure that the test plans, cases, etc. are all inspected by a review team to ensure that other perspectives can be brought to the testing.
It's all relative to the application you are testing and the skill set of the people available for testing. You have to make your own call on that one.
One thing we try to do here though is to rotate testers between projects on a regular basis. At the very least this means a fresh pair of eyes looking over a project. Normally this makes the job a lot more interesting and increases morale as no-one gets bored hanging around on the same project for ever.
It's maybe not appropriate for large projects, but is worth considering if you have a number of ongoing maintenance projects.