The Atex Blog

Once a year I have lunch with Cindy Taibi from The New York Times and her old boss, Stanley Kaplan (the retired systems director, not the standardized test prep guru). It is an enjoyable tradition we have kept up for the nine years since Stan hung ‘em up.

So there we were last week at Ben’s Kosher Deli in the garment district, a few blocks from the Times, enjoying matzo ball soup and hot pastrami and sharing stories of the days when we installed a huge Atex system together.

Stanley is a wise man. He was in systems at NYT from the transition to cold type all the way to pagination and the web. One of the most valuable lessons he ever taught me was the importance of testing. When we were doing the specifications for the J11 editorial system Stanley insisted that NYT users write the test plans. His rationale was NYT people knew how the functionality should behave, who better then to write it up? If his users were responsible for the test results, then they would be motivated to write the best possible test plans. NYT handed Atex test plan after test plan, dozens and dozens of them -- setup/inputs/expected results. It was a great way to communicate to the Atex engineers exactly what the users expected. And it put the users on the hook for quality because they would have to live with the results of the testing.

The old project manager in me cringes when time and again testing time is squeezed because there is a go live target and the set-up work slipped or expanded to encroach on testing. Happy users are a direct result of the quality of the testing. Writing test plans is no fun and executing them is no day at the beach either. But the lesson from Stanley Kaplan, to involve the users in the test plan development, can go a long way to cleaner implementations.