Perhaps part of the reason Keil may be "tired" of the 8051 is that the basic functionality of the architecture has been set in stone for decades. Most variants are supported by simply making a header file, and I think they've wrung all of the performance improvements out of the tools they're ever gonna get.

I couldn't justify paying the maintenance fee for another year's worth of uVision2/C51 support, as v7.whatever I have seems sufficient for everything I do. I don't need the IDE "improvements," nor do I need a simulator that supports variants I don't use.

I suppose Keil does well enough selling new licenses, and they'll probably continue to do so as long as someone's buyin', but I don't know how many users, like myself, don't see the need to pay for maintenance.

As for sdcc: while in theory I agree with everything that's been said about how someone else might pick up the slack if the maintainers get bored or distracted, in reality, how often does that happen?

-a