STAREAST Recap: Share What You’ve Learned
For those who have attended multiple STAR testing conferences over time (I just learned the first STAR conference was held back in 1992!) you’ve probably noticed that while the agendas are always fresh with new topics and speakers, there are a core group of speakers that come to share their knowledge and experience year after year.
Last week’s STAREAST in Orlando, Florida was my fifth STAR event, and over the years, I’ve had the pleasure of interviewing some of the regulars—James Bach, Bob Galen, Johanna Rothman, Dorothy Graham, and others. One of the questions that I’d intended on getting the answer to this year was, “Putting aside your careers as consultants and paid speakers—what keeps you coming back year after year?” I’d intended to get this question answered, but as I attended their various sessions, the answer became evident.
Software testing, agile development, the rise of mobile apps/devices, the next generation enterprise—these issues change, multiply, and present new challenges that many developers and testers may not have faced before, no matter how long they’ve been in the industry. And these speakers, by working with so many different organizations around the globe, pick up on new and emerging trends, methodologies, and practices to tackle these challenges. Thankfully for us, they’re all eager to share them with those who attend in hopes of having their questions answered.
Given that each speaker submits their session topics for approval months in advance, and without any choreography between them, it’s always interesting to look for, and find a prevalent theme and focus, and this year’s was evident from the first day and continued throughout the week.
Customers (and therefore, CEOs) want their software, and expect their software delivered faster than ever before. Oh, and at an incredible level of quality. So, how do we go about doing that?
Thursday’s technical presentation on Thursday attested to this by labeling quality as “paramount,” and that greater test coverage is key to releasing better software faster. Chief strategy officer Wayne Ariola rattled off shocking numbers around software failures by financial institutions, and the well-publicized multiple outages of the Sony Playstation Network after a security breach that cost the company $18 billion dollars.
Ariola demonstrated Parasoft’s service virtualization offering as a method for “simulating test environments to help development and testers with earlier, faster, and more complete testing.” The “test early, test often” practice was echoed by many throughout the week as an absolute must for any team striving to eliminate the high cost of bugs not being found until reaching production.
Earlier on Thursday, voke market analyst Theresa Lanowitz discussed the trend of software failures dominating media headlines and the subsequent call for “better testing” even by president Obama himself after the abysmal launch of Healthcare.gov.
Lanowitz pointed to high quality tooling available today, with it’s low cost, and low barriers to entry as something testers, and entire organizations need to look toward as possible solutions for removing the barriers and inefficiencies that have plagued software development for years.
Her keynote on “extreme automation” focused on companies like Skytap, who she directly named, and our enterprise-level product built to eliminate IT sprawl, reduce hardware costs and setup time, and empower development and test teams with what we call environments-as-a-service (EaaS.) These are environments that can be quickly spun up and down, cloned, shared, and fitted with continuous integration tools like Jenkins, and are production-grade for easier deployment—even on AWS infrastructure.
I highly recommend viewing Theresa’s keynote if you weren’t lucky enough to attend STAREAST in person or virtually.
From a non-tool-centric approach to delivering better software faster, SmartBear chief performance evangelist Scott Barber gave a great session, titled, “Application Performance Testing: A Simplified Universal Approach.” I loved Barber’s definition of a “performant” system:
A performant system is optimized to balance cost, time to market, and capacity…while remaining focused on quality of service to system users.
Barber humorously, yet oftentimes accurately, cited that your CEO isn’t likely to ever ask you, “How good is it?” when referring to the product you’re testing, and much more likely to ask, “When is it going to market and when are we going to start money.”
His solution revolved around getting buy-in from developers to check their code’s performance every day by adding just a small bit of code to their work. By exporting the performance/refactoring results to a CSV file, teams should be able to produce visible trends that can catch potentially catastrophic performance issues at their inception, not waiting until production to find them, where everything comes to a screeching halt. There’s that “test early, test often” mantra I told you popped up everywhere.
My last session of the week was a presentation by IBM’s Allan Wagner, who also spoke to the power of service virtualization in order to achieve continuous testing, and therefore release higher quality software sooner. I loved Allan’s presentation, and the survey he quoted produced almost identical results to Skytap’s own findings through extensive research in the testing community.
Those who would like to participate in our research can do so here, and you’ll receive a free copy of the results in a report to be released soon!
Wagner stated the obvious when he remarked that many testers simply don’t have time to not only do the work they love to do, nor can they provide the test coverage their projects absolutely need to ensure high quality. But it was the why don’t testers have time that was both fascinating and disappointing. The number one response: “We’re spending too much time setting up and configuring automated testing tools.”
Wagner’s solution was similar to that of Lanowitz, Ariola, and Skytap, stating not just that, “Testers need production-like environments which are realistic and complete, available and on-demand, and inexpensive to setup,” he also reminded everyone that these types of environments do exist and can make a difference in your dev/test efforts immediately.
Wagner went on to point out that the best offerings out there today allow “developers and testers to continue to use their current testing procedures and tools” as well as the “streamlined creation of production-like environments” and that by not taking advantage of these types of environments you’re allowing bottlenecks to continue to handicap your testers by forcing them to defer their testing efforts until far too late into the SDLC.
As I made one last walkthrough the STAREAST expo where many congregated before heading home to their own projects, eager to try out many of the ideas presented to them over the last four days, I overheard someone say to their friend, “we’ve got a lot to share when we get back.”
While the challenges testers are dealing with may change, this is the overarching theme of the STAR conferences. You’re going to take in a ton of information in those few days, and some of it may work for you, and some may not. The point is to go back home with some fresh ideas that don’t just benefit you and your team; they were shared with you to help you entire organization get better, and continue to get better year after year.