I have about 35 years experience of assembler programming, and have been professionally using C and its derivatives since about 1980. My personal preferences now are to use whatever is the easiest to get the job done. Clearly in many environments a High level language is unquestionably the most productive, but when the need is to get down to bit twiddling and squeezing code in small amounts of memory then it has to be assembler - or at least something where you can have clear transparency of what is going on under the covers and the means to control it.

I'm probably at my happiest when I can use a compiler to do most of the grunt work followed by grabbing the generated assembler and tweeking if necessary - but visibility of what is going on is the most important issue. Keil can do that but it took a lot of digging to find out how.

I inherited the project I am currently working on, one part of it is about 4k bytes, and running out of rom space - all in C, but about 600 bytes was a library routine called in to generate a random number - it crunched an IP address to generate a 4 bit offset used only once - just shows what can be lurk under the hood of RAD.

Rod.