When to Say NO

Here is a good story for you. Tell me if what you will do the same thing or would have done it differently.

Day 1 of 15. Your team was over capacity. Day 6 of 15 business say that they want some new investigation done. Really important to get done. This was brought in so that we can fix the problem next iteration. So this was all fine by me as this does not affect QA. I am happy. We are still at capacity. Then day 8 of 15, they suddenly make this investigation story a fix it up story. I was asked if we can do it? He even dangled that the developers will do an automated integration test. I said if I will be satisfied with the test then I have no problem with it.

I then asked the manager does he know the complexity of the change. He said he really did not know. So that time I said NO. Main reason why I said no is was this change localized to a component or will it be for all components? I would not know how much testing I needed to do. And time was running out.

Next day, day 9 of 15. I was able to talk to more informed people such as the developer. They knew better on the scope of change. We discussed on how big the changes were. One of the change was fairly localized to one component. The second change was more tricky. It was updating an underlying the change. It will definitely affect 2 of the components they just have to be careful in updating the project so that they only update the 2 components and not the rest. So armed with the knowledge of only needing to test 2 rather that 8 processes, I said lets go ahead with the change.

So developers continue with their development. And they are done and code checked in on Day 11 of 15. So we can now start testing. Problem happened, this information was not disseminated to the rest of the development team. so guess what happened? Code was built but SCM updated all projects with the configuration that we said should not be applied to all projects.

So we talked to SCM and asked if we can revert the configurations back. At first SCM was saying it was only a configuration change that was changed. No functional change. So as QA I should not worry about the other processes. My QA instinct was just saying that I should not just release something that I am not worried about. I cannot take their word for it. I have to prove it. So I kindly said NO again. I just won't let a piece of code go out in the wild with out testing. So we asked him if he can revert.

He tried to revert it back. And .....POW! Another problem blocks the road. There was another team touching the same code base. And they have updated the parent project. So now it won't be easier to revert the changes. So what are my options at this point?? Well, I guess I have no more options. I am stuck with this configuration. I have to do a regression on the rest of the components.

So what was scheduled to be a quick 1 day test turned out to be a 2.5 day testing effort. and we were not done till day 14 of 15.

Should I just have stayed with my first NO? Was there a place that I have done better?

I honestly don't know. We will find out after we release the code. If we do solve problems ( Major problems ) then I would say what we did was the right thing to do. But if the changes does not do anything, then I would have say we need to look back.

What is in your toolbox??

I often get the questions by my peers on what do I use for different kinds of QA related activities. So here are a few tools that I use. Hope this helps you.

Test Prep
During test preparation I often use CTE XL. This is Classification Tree Tool. This is a nice and simple way to lay out all your objects and classes and easily come up with your test cases. It can export to different format. I usually just export the test cases to an excel spreadsheet.

Another way to prepare a test is by doing a mind map. A good free tool to use is FreeMind.This is free easy tool to do your mind mapping documents.

In testing, I am usually face to creating data that I need to enter in the systems. And at times, I feel like I keep using the data over and over and over. So a tool that I found that helped me with generating test data is generatedata. Before you can use this, you need to set up a couple of stuff before hand.

Testing
Some quick automation tool, I use WatiJ(pronounced as wattage). It stands for Web Application Testing In Java. I don't have a lot of experience using java, but I found it easy enough to use referring to te API documentation. This needs a java ide. I really like IntelliJ and Eclipse. IntelliJ is not a free application while Eclipse is a Free application. Watij is a free software, so help is a little bit limited. There is a good user group that is really helpful.

Another good automated testing tool is QuickTest Pro from HP. This is a commercial so support is really good. I recommend this software to 2 of the companies i had work for. And for both times, it has really worked great. I would not recommend this for smaller companies as it can be expensive if not use wisely.

Another free tool that I use is Fitnesse. Now this tool is a bit tricky. Cause the whole organization should commit to using this to be successful. This needs some time from the development team as well to do the the Fitnesse Fixtures in the code base.

For testing the UI of a web page, I often use IETester. This allows you to switch easily from different Internet Explorer versions.

Some outcomes for testing is a big text file. And often, to test the outcome is to compare 2 text files. A good tool to use is WinMerge.