Over time most embedded design people reach familiarity with a controller and mostly C language. And with time we do get better performance and more functionality from the silicon suppliers. I work mostly with the LPC (ARM7) family and I am happy with it.
I spend 3 hours reading a short article and looking at the tools of the new Propeller Chip. Parallax designed a simple 32 bit CPU and linked 8 units in one silicon solution. They kept the approach simple and with a round robin type the8 controller get severed sequentially. Looking at the new language SPIN they introduced (mix type of language I see object, pointers, special instructions) and an assembly language.
I am not so sure what I should take of this approach. I don't see it as an incremental improvement process here, but rather a complete new embedded concept. I read that the SPI interface is missing and it is a drawback, but I am sure this and other shortcomings will be soon addressed.
What I am not clear is how a faster design task e.g. serving IO pins in few hundred ns range would be implemented. Because, if the design often can only be implemented in their assembler language I see this as a real drawback. Also, because it is not C language all the legacy software needs to be rewritten! Big issue.
The whole controller concept is proprietary to Parallax and no real competition will be available. The IC is $25 single units and not much below $20 in volume and you still have no analog capability. You do get really good silicon for this kind of money on other embedded CPU concepts.
I give Parallax a lot of credit for their work & design and I don't intend to badmouth here. It is just that I am currently not convinced that it is a long term viable solution. Although reading their MIPS it is impressive. Will this concept fly or soon or later end in no where? Thank you ..richard