I certainly understand the position of a computer user who sees and suffers from the mutual "finger pointing" that occurs when a new operating system version is released and suddenly applications that were working are no longer fully functional or exhibit odd mannerisms under the new operating system software. Also, given my status as an employee of Adobe, I expect that you will certainly assume that I have parochial biases on this topic. But here goes ...
Over my long career I have not only been involved with the design, implementation, and validation of application programs, but also managed groups that designed and implemented operating systems as well as system software used in conjunction with same. In those roles, one of the cardinal rules that we followed was that new operating system versions and features were absolutely required to maintain full compatibility with existing application programs. No ifs, ands, or buts about this rule. Just because a new feature was neat, cool, ... cool, neat did not excuse the OS developer from maintaining application compatibility. Application compatibility testing was at least important as testing of new features. Anyone in my organization who knowingly violated this compatibility rule or didn't immediately resolve any such problem unknowingly introduced during development and found during our rigorous testing risked termination of their employment with our company. And I know similar attitudes and practices existed with other mainframe and minicomputer manufacturers (I was with Wang Labs at the time).
Quite frankly, it is (or at least was) very reasonable for an application program developer to assume that unless their program or installer does something very stupid (such as refusing to install or run after interogating the OS version and finding a newer version than expected), that such software should continue to run as it did previously, although perhaps not being able to take advantage of newer OS features.
Apparently, today's OS developers don't really value the stability of their operating system environments for existing applications. That's exactly what we are finding. Neat, cool, ... cool, neat eye candy seems to trump operational usefullness and compatibility.
Over the history of MacOS, forgetting for the moment the extreme costs incurred in the conversion from Motorola 68xxx processors to the PowerPC processors and then from the PowerPC to the Intel processors as well as from MacOS 9 to MacOS 10, Adobe has made tremendous investments in reprogramming and testing to simply assure that the current and next application version works on the newest MacOS dot release. Such investments have come at the expense of new feature development. Apple doesn't reimburse us for such reprogramming or the extensive compatibility testing we do for them.
In this particular case, we did not say that CS3 won't work on MacOS 10.6, but rather, we haven't officially tested it. We did extensive compatibility testing with CS4 and new versions of our software under development. Hopefully the fixing of buckets of bugs that Apple fixed as a result of such testing also help CS3 run under MacOS 10.6. It is also fairly unfortunate that Apple does not have an extensive end user beta testing program in addition to its fairly limited developer testing program. Such end user testing would better find application compatibility problems with legacy versions of applications used in the field and allow users to put pressure on Apple to fix such issues.
- Dov
