Software quality attributes and their rankings

How much do Agile techniques, Agile methodologies, automation, certifications, and a formal QA team affect quality? In this second part of a three-part interview, we explore some of the 121 software attributes ranked by quality value.

“The rankings come from observations in about 600 companies and 13,000 projects,” answered Capers Jones and Olivier Bonsignour when questioned about the table in their book, The Economics of Software Quality. The table lists 121 software quality attributes and their rankings on a scale from +10 (extremely valuable) to -10 (extremely harmful). In part one of this series, we learn that use of Agile methodologies and hybrid methodologies...

are both ranked as 9.00, while use of the waterfall methodology is ranked as 1.00. In this second part of the three-part series, we explore more of the attributes and their rankings.

SSQ: Though “Use of Agile methods” ranked a “9.00,” I didn’t see line items for many techniques that some Agile teams practice such as test-driven development, pair programming and continuous integration. Have there been studies to demonstrate which Agile methods are the most beneficial in achieving higher quality?

Capers Jones/Olivier Bonsignour: One caveat is that Agile is not the only method that is effective: both the Team Software Process (TSP) and the Rational Unified Process (RUP) are also successful. While we’ve been able to get some measurement regarding Agile methods overall, the industry around Agile is not very data-rich. We have not seen much activity in the Agile community towards quality measurement programs, especially at a level of refinement that would provide insights into such questions as variation between Agile methods. When we had our research paper on technical debt measurement accepted at this year’s 10th anniversary Agile conference, the organizers struggled to figure out where to feature the talk, since there was no measurement track.

SSQ: There also aren’t line items for specific Agile methodologies such as Scrum and XP or mention of Lean or Kanban. What are your thoughts on the different Agile methodologies?

Jones/Bonsignour: It’s probably too early to derive any conclusion. Most of the Agile methodologies are less than 10 years old, and moreover, most of them have been built on the ground defined by the previous ones. There are interesting points in all the different approaches, and most of the time it’s quite difficult to fully embrace one and only one methodology. Most of the companies we’ve been talking with are in fact using a sort of mix of different methodologies picking the appropriate points/parts in each of them. For sure, with Scrum currently being so popular as to often be synonymous with Agile, you will hear most companies say that they are using/implementing Scrum. But XP includes more technical practices and Kanban/Lean are nicely complementing the good ideas/common sense behavior promoted by the above. Lean is not so much a method as an approach to focus on eliminating waste from the process. Lean tends to be a broader management initiative to introduce measurements to identify areas where cycle time is being degraded, identify root causes, and then measure again in a cycle of continuous improvement. As such, Structural Quality is an important component of Lean application management.

But again, we are not aware of any definitive studies of their relative effectiveness. The question here is whether we should even look for one. Aren’t the principal characteristics of Agile the agility and flexibility? So shouldn’t the Agile practitioners be Agile enough to not stick to a single approach and instead continue to do what the community has been doing for 10 years: quickly adopt the new improvements?

SSQ: “Use of automated test tools” scored an 8.00, but there are many different kinds of automated test tools. Which type was this referring to and what were the pros and cons?

Jones/Bonsignour: For a number of reasons it is not possible to name specific tools.  In the case of automated testing there are a number of vendors and a number of tools available. Tools that run test scripts, defect tracking tools, test library control tools, and several other categories are in this class.

SSQ: Line items 109 and 110, “Certification of test personnel” and “Certification of SQA personnel” are both ranked as 8.0. What kind of certification are we talking about? There are some that would argue that hiring managers depend too heavily on certifications. What are your thoughts on that?

Jones/Bonsignour: As we all know, the software industry does not yet use licensing or board specialties as does the fields of medicine and law. However, there is evidence that test and quality personnel who care enough about their work to take courses and pass certification exams from either non-profit groups or companies that provide such training and certifications have overall levels of defect removal efficiency that are higher than equivalent projects with untrained and uncertified personnel.

SSQ: Line item 47 ranks “Use of a formal SQA team as a 9.0,” yet many who practice Agile discourage formal teams organized by domain area, due to the organizational silos that can develop. What are the reasons behind a formal SQA team enhancing quality?

Jones/Bonsignour: If you look at the nature of Agile development, it is aimed primarily at small projects below 1000 function points where total team size is less than a dozen people and they are all located close enough to have Scrum sessions and frequent meetings. When you deal with projects large enough to have teams of 500 people scattered across half a dozen countries, you need the consistency provided by a software quality assurance group. Below 1000 function points in size, quality assurance teams seldom form. Above 10,000 function points in size, quality assurance teams are much more common.

In addition to a formal SQA group, if the team is enhancing or maintaining an application larger than 1000 function points, Structural Quality issues also begin to surface. One change to an application that’s large and interconnected can cause unexpected behavior in ways that are hard to predict without analyzing the structure of that application with each build or release cycle.

For more of this interview see:

Quality metrics: The economics of software quality – Part one

Quality metrics: Changes in the way we measure quality – Part three

 

This Q&A is based on The Economics of Software Quality (Jones/Bonsignour) Addison-Wesley Professional, and can be purchased by going to http://www.informit.com/title/0132582201

For a comprehensive resource on measuring quality, see Quality metrics: A guide to measuring software quality.

Dig deeper on Software Quality Management

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchSOA

TheServerSide

SearchCloudApplications

SearchAWS

SearchBusinessAnalytics

SearchFinancialApplications

SearchHealthIT

Close