On the 8051 I like the wide range of applications, which I can make with using the same knowledge, experience and development tools.
E.g. for small, fast and cheap applications I use the AT89C2051.

And for bigger things the DS80C320, T89C51RD2 and many others available and still not too expensive.

Also the wide range of peripherals unbeaten on the 8051. E.g. the most devices with built in CAN are 8051 derivates. And Dallas announced a new 8051 with built in Ethernet (DS80C400).

And because the wide range of manufacturers I have never big delivering problems.


Also I can absolutely not agree that it was slow. Since the most potential to write slow or fast programs based on the programming style, not on the real speed of the CPU. Nevertheless, if more speed was needed later, also high speed derivates (DS89C420 = 40MIPS) exist.
Also the interrupt logic with 2 or 4 interrupt levels can make it easier to implement real time applications. E.g. on such micros with only one interrupt priority a second priority must be build fully in software and then the speed was slow down dramatically.

Also powerful and well tried development tools exist, like very efficient C compilers. Thus C can be used on the small AT89C2051 and also very big applications fits in the standard 64kB linear address space. Again, if more needed, then derivates above 64kB exist.


... and many other reasons make the decision to the 8051 family to a very good choice.


Maybe for very high volume applications the PIC can be a little cheaper and thus the longer software developing time because its awkward architecture can be tolerated.
But if than changes need a bigger derivate, its very complicated to reuse the older software, since on bigger derivates big changes in the architecture (12-bit, 14-bit, 16-bit instructions) cause a different instruction set and memory mapping.


Peter