Rough Stuff: a difficult day at the testing center

Not a good day on the testing front….

June’s caught the flu, so after some back-and-forth we decided Sarah and I would attempt to complete the tests. So we headed downtown to Tony’s testing center. Since Tony and I have differing test objectives, he and I had a short chat about the day’s strategy, and I adjusted my plan a touch to accommodate Tony’s need to be able to report progress on a specific problem. Then Tony joined Margie as she and her boss, James, headed off to demonstrate the web application for a customer group.

Sarah and I began by exploring the scope of a problem June had identified on Monday. There’s a problem with the program flow; under some circumstances, the app skips a couple pages, resulting in incompletely filled forms. We verified the problem, and identified the circumstances which trigger it. We also found a work-around, but it’s so ugly we can’t call it acceptable. I started to document this, but my remote e-mail access collapsed before I could complete it. This is now on tomorrow’s To Do list.

Tony’s main concern is that I test the mechanism his staff has devised for delivering Adobe Acrobat files to our customers. The delivery system has been completely rewritten since the last test session, so the opportunities for failure are numerous. The revised system changes the delivery channel, adds email notifications as the work becomes available for pickup, and has some technical changes intended to address security concerns. If the changes work, the system will deliver fairly large query results to our users in Acrobat format, while intercepting the truly large files for delivery by alternative methods. We need to ascertain that the system can actually deliver the large files, that the e-mail notification works properly, and that it can differentiate between the merely large files we want delivered and the truly enormous files we need to intercept and send to the printer.

Sarah had some appropriate test cases identified; I had a different set. The first-step plan was to input our queries, wait for them to process, inspect the results, and decide the second step. Unfortunately, our first couple cases served to demonstrate a glitch; the system was clearly failing at the exact point the back end application notifies the web app that the file’s available. I called Jamie, our technical support guy, and asked that he stop, then restart, the process which seemed the likely culprit. The fix worked. Back to testing.

The next few tests demonstrated that the system had lost its connection with FileNet; this broke the tests, since it is necessary to combine many smaller documents to create the large documents we needed for testing. I had some notion of the cause, and could have done some investigation had I been at my own desk. Or I could have called Margie, who’d have checked from her desk except she was also out of the office. I tried to contact Jamie, but he wasn’t readily available. Eventually I called Steve, who used his hard-won VPN connection to rebuild the FileNet interface. By noon, he had it working.

Sarah and I went to lunch. I had to be at the big call center meeting, so we didn’t return in the afternoon.

Ugly. Tony’s not happy, I’m not happy, no one’s happy.