Despite the name, most of QA groups/departments in the software industry are not QA departments at all. They are a testing group that design, implements and execute test plans, test cases and test scripts. Nothing wrong with that indeed!. However, the next step to improve the quality of the software is to go beyond the testing phase and create a culture of quality among all the participants of the software development process.

How is possible to begin this culture of quality? Ideally you will implement any of the existing models (CMMI, ISO...), however, given the amount of effort and time (so, money!) involved, it is usually difficult to get committed support from senior management. A group that incidentally looks at testing and QA as non-core/non-fundamental software development process activity.

Here at Visiphor the Test Group is reinventing itself into a QA group by pushing the following list of actions that are easy to implement. Incremental change is powerful if steady. You do not need to apply all the items, you can pick and choose or add your own. However consistency is important. If you pick one of the items, you need to keep it during all the development cycle: it doesn't make sense to peer review version 1, 2 and 4 of a document but not versions 3, 5 and 6 of the same document!

Here is the list. Comments, suggestions and real life stories gladly accepted. I would love to read your comments and what has worked (or not!) for you.

Action	Description	Who is responsible?	Phase
Peer review of documents (architects)	Each document is reviewed at least by another member of the Architects team Architects	Architects	Planning
Peer review of documents (developers)	Each document is reviewed at least by another member of the developers team	Development	Planning, Development
Peer review of documents (tester)	Each document is reviewed at least by another member of the testing team	Testing	Planning, development, stabilization
Peer review of documents (systems)	Each document is reviewed at least by another member of the systems engineer team	Sys. Eng.	Planning, development, stabilization, deployment
Solution Setup (developers)	A senior developer creates the Visual Studio solution(s), project structure, and namespace framework.	Development	Planning or Development
Source Control Repository Setup	A senior developer adds the solution(s) to the source control repository.	Development	Development
Code Reviews (developers)	Similar to the peer reviews of documents, the code for the application is regularly reviewed by another member of the team or by an automatic tool.	Development	Development
Kick off meeting. Also known as Launch Meeting.	Everybody involved in the project needs to be present. The purpose of this meeting is to fulfill several of the MSF principles of common goals, vision, foster communication….	PM	Planning
Setup and use of Issue tracking system	An Issue Tracking System needs to be set up for every project, in addition, it needs to be the central collaboration tool that each team member uses to identify, record, triage and solve bugs and change requests.	Testing	Planning
Sign on at each iteration/Marking Iteration boundaries	The beginning of each new iteration (as defined in the MSF process) should be marked by a team meeting with the same purpose as the Launch meeting. Iteration ends should at least be marked by an email from the PM, and in some cases by a post-mortem-like team meeting. The goal is to keep everybody informed, and with the same objectives.	PM	Through the lifecycle of the process
Dedicated Test environment	The Test environment should be used exclusively by testers and not shared with development.	PM/Sys Eng/Testing	Development, stabilization
Unit test	Developers, should unit test each piece of their code either manually or using automated tools	Development	Development, stabilization
Define the Scope for every project.	It is important to define what is in the project and what is out. Gray areas will always exist but need to be minimized.	PM, Development, Testing	Envisioning
Customer involvement in the testing effort of User Interfaces	The rationale of this point is based in the premise the customer is the best to judge the user interface: both the information contained, the user-friendliness, the work flow and the business logic behind it . In this context, if the customer is involved in the testing, the quality of the end product will be better.	PM	Stabilization, deployment
Project will proceed in a continuous stretch with not time gaps in between.	When a project stops, and then restarts, a lot of the knowledge accumulated is lost. Projects done as a continuum will have better chances to reflect better the customer original intentions.	PM	Through the lifecycle of the process
Post-mortem meetings a the end of each iteration	In a multi-phase or multi-iteration project, this practice should help to improve process and correct problems that have appear in previous phases	PM	End of the project
Setup and use of Change Request Procedure	The team, PM and customer should set up a procedure to deal with new/changed requirements. I should include a risk assessment for the requests.	Testing, everybody	Planning
Setup and use of Requirements System	A centralized requirements repository will reduce the dispersion of information and ensure all the member of the team share a common list of requirements. It may allow the possibility to generate metrics and be able to show to the customer all the requirements changes.	BA, Testing, PM	Planning
Data dictionary	A spreadsheet or document that describes the name, meaning, type and other information about the fields of a database and how they transform (if applicable, as not all the projects may recommend a data dictionary)	BA, Development, testing	Planning, Development
Monitoring of the systems for unauthorized changes	This monitoring should be associated with clearly defined consequences for intentional unauthorized changes	Everybody	All phases

Soon I will post a second list of recommendations that are a little bit harder to implement or less applicable to all projects.

Note: The "Phase" is in the context of the Microsoft Solutions Framework (MSF), and are orientative. Some of the tasks do not have a clearly defined phase in which they need to be done.

I have read the interesting article about Michael Porter's ideas titled 'Why Do Good Managers Set Bad Strategies' (
http://knowledge.wharton.upenn.edu/article.cfm?articleid=1594). The answers are interesting, although I have some concerns about the premise: Excellent managers can be very bad strategist and vice-versa (In WWI Germany was very well managed, but its long-term strategy proved disastrous, while the new and chaotic Soviet Union became a World power).

However, it is his statement of "Strategy has to do with what will make you unique" what make me scratch my head. In the current context of the software development industry, specifically in the context of a small professional services firm like the one I am working on, what we can bring that is unique?

Perhaps it is a self-interested answer. However, it is clear for me that currently that major differentiator among our competitors is Quality, used in the broadest possible terms. Quality understood as the set of processes that will be able to deliver in time in budget a solution a business need in accordance to the customers' expectations. Notice this would include managing the customer's expectations!

After more than ten years in the software industry, still surprises me that so few are trying to exploit this niche as differentiator. It still surprises me that many managers, visionaries, captains of industry and so on see QA as a luxury or fat to trim. In most of the cases, it is a simple marketing investment.

In an ad from Mercury, they say, "80% of applications are deployed untested. 100% of customers really hate that.” So, if this is the state of the industry, let me return to my suggestion and to the point of this lucubration: what about developing a strategy that will focus in the QA of your product as a way to make it unique?

The last but not the least. The article also includes the quote "If you don't pursue a direction for two or three years, it's meaningless." Putting the quote in the QA context: you should not waste energy and efforts implementing a lukewarm QA strategy. Do not set a policy and then constantly break it. For QA professionals it is frustrating and you will not achieve any good.

I have been reading the massive article

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnpag2/html/mtf.asp

posted in the Microsoft web site with the title of "Testing .NET Application Blocks–Version 1.0".

I can not say that I liked it. I consider it useful, I think that will take me some time to fully digest it contents, however, I would say that it could be written in a less bone-dry tone. For example, I should be excited to read the article after reading the introduction, not to consider that it could be appropriate to read it...

In addition, there are tendentious comments :-). For example
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnpag2/html/mtf_ch02.asp
"Testing Methodologies" is not an honest report about the different testing methodologies. It is a naked attempt of Agile apologetics. After the authors, Agile has no disadvantage at all! What about scalability for example? The more I know about Agile, the more attractive I find some aspects. However, you should not forget all the strings this methodology has attached: a competent, committed and engaged customer for starters....

In any case, if you have lots of patience, are not sleep deprived and have some time, it is a reading that I recommend.