The thing that started OOPSLA was a small number of pioneers in a new-ish software technology - Object-oriented programming. It was a hot topic among academic researchers, a variety of companies were looking at OO as a new [practical?] way to build products. Places like Bell Labs (one I'm familiar with) were busy trying to figure out a way to turn OO programming into a competitive advantage.
As more things happened, OO turned from researchy (and riskyscarynot ready for prime-time) to novel. The distinction I'm trying to make is that non-researchers were using OO approaches and discussing them, promoting OO as a viable approach to building real things. That lead to 2 classes of audience at OOPSLA: people pushing the front edge of the technology, and people looking to use the trailing edge [working?] technologies to do something.
This latter group forms the bulk of the current OOPSLA body. They want to come an hear from the experts. Now that Objects are Mainstream, and perhaps this echoes JavaKilledOOPSLA, there are many other places to get this kind of information - JavaOne, SIGS conferences, JAOO, .... The only advantages OOPSLA has over these other places are
The problem is that because objects have succeeded, the other meetings, better focused on training people on how to use objects, also meet #1, and these people don't care about #2. --RonCrocker 26JUN04