September 23

I was up first thing this morning and ready to get right back down to work. I was drinking my coffee, starting up the computer and as the desktop comes up an alert that there is a system update available. The easiest thing in the world is to say install. It turns out that it would have been much easier and certainly smarter to wait until the reviews come in and I know its safe to install. I had to restart after the install and a hard crash occurred, a kernel panic. This is the worst type of crash that can occur. It offers no choices but to shut the computer down and restart. It only took a couple of restarts to identify the MOTU 828 firewire interface as instigating the problem; If not plugged in and turned on the computer started up just fine. I spent most of the morning monitoring macintosh websites and a dedicated mailing list for Digital Performer. The truth is I could easily have avoided the whole problem if I was not in such a hurry to get down to work. There were a couple of posts to the list regarding possible problems with the update whose titles just did not register with me. it was not clear whose problem this would be to solve ultimately, but by Noon it was clear that I was going to have to find a workaround. I spent the rest of the afternoon installing a new OSX system to a different drive partition which had been dedicated to OS9 but still had room enough for a basic OSX install. Then I updated to 10.2.6 which was the last known good version and began to figure out how to reinstall drivers and other software to support the mixing environment I have been working in. By 5 PM I was finally able to open song files without errors occurring.

The good news is that I was able to create a mix of Little Girl in this new configuration and save it. The bad news is that this new setup is fairly shaky and Performer has crashed several times. The last one was after I had just finished recording the vocal for Wayside Hotel and which was then lost having not been saved before the crash. Yet another minor disaster that I could have prevented by practicing good habits that i have advised others to use for years. I managed to stay pretty calm throughout the day though there were definitely a couple of times that I cursed Apple for this even though as I said I could have prevented it.

These events bring me back to the time when I worked at Apple as a software tester. I can remember tremendous arguments over what testing would call a show-stopper bug, usually a crash that was found just before a system rev was to be shipped. One arm of the company always wanted to ship the product and fix it in an update and testing would do everything in their power to prevent the end user from having to suffer the consequences. What usually happened was that the argument determined the true level of importance of the problem in the face of the need for the company to ship products and usually the right decision was made, sometimes it shipped and sometimes it was delayed. I cannot remember ever seeing an update go out with as many serious flaws as the one Apple released today. The update was pulled off the servers by 3 or 4 PM. I would guess that their tech support lines were flooded with calls and someone with clout was finally convinced that more harm than good was being done and it would only cost them a lot more money to allow people to continue to download the update. When I left Apple I was witness to a fundamental change in the way the products were being tested. More and more of the responsibility for testing was being placed on the engineers themselves and there were fewer and fewer testers on any one project. With the advent of OSX there also seemed to be basic changes in the way that engineers were able to check their code in and out. These changes seemed to make it easier for mistakes in generating builds to occur. What I believe happens is cyclical in nature. It is like the tide where forces exert pressure which over time sways the momentum from one philosophy to the other and then back. The engineers versus the testers and who gets to call the shots.It is inherent in the relationship that engineers feel superior and believe that their opinions are the only ones that count. Testers believe that engineers are blind to their own mistakes and engaged in denial when a mistake is reported. So the extreme case occurred today where a new build was foist upon the unsuspecting public, myself included. I truly hope that the tide has reversed now and that more emphasis and resources will now be devoted to testing. If I am wrong about the level of commitment to test organizations at Apple then I hope they will lose seem dead weight and go out and retain some better testers. the one thing that is clear today is that they could have easily found most of the major problems with todays build if they had looked a little harder. If they are having trouble finding experienced Mac testers I know of at least 2 who are out of work and can start right away....