I not only changed from the 65C02 to a range of others, e.g. 805x, 6805 and later the HC version, 6801/03, HC11, and an occasional "other" (TMS320-series). The 65C02 always had the instruction set I preferred, but since it doesn't exist, having been killed by the later Commodore video toys, et al, in favor of the more highly integrated parts, it's just not on the menu any more.

Now it's HC08 and SCENIX, I guess.

I've started looking at the MCU's more as logic components than as computers, however. I find the fact that several manufacturers make small package, reduced-resource parts an invitation to partitioning tasks in a way that acutally speeds up the development process, allowing the team leader to parcel out work based on firm interface control spec's and treating each team member's work as a separate device. The small devices e.g. Motorola 68HC908K, and the like, really do invite one to do this. With the HC05's it was even more pronounced, as the '705kj cost only a dollar in modest quantities, yet was quite capable. If you put several of them to work rather than using a single larger-poin-count part, the partitioning was impossible because the device had to share pins between the various tasks.

Uli