I do not say that this is a wrong way, but I see this strange. With such way in the future we will have fantastic hardware and kewl hardware ingineers but what is about software programmers?

Oleg,
I can not decipher the "kewl" mistype, anyhow I understand your point.

The progression towards software made by methods compatible with the uninformed is not as much an engineering issue as it is a "time to market" issue.

I have been involved with a few large scale developments and to meet the time to market, you have to work with what you can get. It is, unfortunately, impossible to find 8 capable developers familiar with the '51, so you end up with one or two such, "assisted" by 12 "I know C" types.

With the education provided today "I know C" types will soon be a thing of the past and by then the next generation will have to work with "I know xxx' types.

As far as hardware goes, the writing on the wall is that "the PCB" will be inside a FPGA/CPLD and the Triscend approach will either reemerge or be superseeded by "code that make a processor".

The problem with this is that we will, some day in the future, have a bunch of engineers fluent in VHDL (or whatever the hardware langue will be at such time) that have absolutely no idea of how to connect their excellen FPGA to the outside world.

Maybe I am too pessimistic, but this is my view.

Erik