Testing continues. All but one bug is fixed, and I’m in a frenzy of regression testing.

Let me restate that. I’m in a very slow, viscous, sticky frenzy of regression testing. The energy is frantic, but what I’m actually doing is remarkably little, and even that is few and far between. I start up a test from one of dozens of application tests we have. These are designed to exercise specific features of the product, ensuring they work properly and fail gracefully when mishandled. I then wait, and then act once it’s finished, which can take anywhere from one to twenty minutes. This gives me time to sit at my desk and ponder, or catch up on paperwork, or something similar.

The problem I have generally with testing in this product is that often I’m called on to test things that I myself have written. This is akin to an author editing his own novel; sure, she’ll do a good job, but the author sees patterns, sees things in terms of chapters and paragraphs rather than in a string of words, and will overlook things that a trained editor won’t. Fortunately, regression testing is cold and distant and doesn’t involve poring over details with a fine toothed comb.

Not glamorous, testing, but very important. Fear of regression runs high here; the feeling is that at any moment, the slightest change could cause a rift in space-time. If we do not rigorously test everything, the terrorists have already won.