STPCon 2007: Day One
October 2, 2007 – 6:06 pmToday was the first day of the three day Software Test & Performance Conference, held in Boston, MA (well, Cambridge, actually). On schedule for Day 1 are several full-day tutorials. I sat in on Scott Barber’s “Performance Testing Secrets in Context.”
This class covered the material in the recent book (currently available only in electronic form) Performance Testing Guidance for Web Applications, of which Scott Barber was a contributing author.
In general, the class and the book do a very good job of detailing performance testing methodology and explores the “core activities” that are done for all successful performance testing initiatives. Having read the book prior to the class, there was very little new in the class that wasn’t covered in the book, but going through the exercises reinforced the concepts, and the lecture/class discussion was interspersed with entertaining “war stories” from Barber’s wide variety of experiences as a consultant.
The biggest takeaways I took from the class were some of the areas where I could challenge the status quo at my own company. For example, not committing to unpredictable test schedules and project plans set by “men in ties,” and challenging some of the performance “requirements” in place at my company.
Some of the other interesting things Barber said (I’m paraphrasing here, these are not direct quotes):
- I don’t create project plans, give me the project plan, and I’ll fit my testing into it
- I don’t create “test plans,” I create “test strategies”
- There is no such thing as concurrency, but focus instead on activity per unit time
The material is presented (by Barber’s own admission) with a consultant’s bias. What I mean by that is there are some of the things that Barber can likely get away with being a consultant, but won’t necessarily fly for a team lead trying to justify headcount and software licenses. There are differences between completing a performance testing consulting engagement and building a continuous team and processes.
There were two topics that were discussed briefly during the class, but aren’t touch on in the book: Continuous Performance Testing (as part of development and/or process) — or, more generally, when should performance testing be done? — and client-side performance testing. Perhaps by the end of the conference I’ll have some information to update about those areas.
That’s all for now, check back in tomorrow for a write up of the keynote speech and the 4 technical classes that I’ll be attending.