Compiler design used to be considered very complicated. While it still isn't the easiest subject, it's been very well analyzed and researched and is now taught in many universities. A semester project often is to write a compiler--granted, it's not a full-fledged optimized compiled...

I would recommend "Compiers: Principles, Techniques and Tools" by Alfred V. Aho (The "Dragon" book, as it is fondly called in compiler development circles, due to the graphic on its cover). This is one of the most popular books that introduces you to the subject.

Also, "Modern Compiler Implementation in C" by Andrew W. Appel is a decent book regarding the issues surrounding optimization.

Both books are avaialble at Amazon.com.

Also, check out "yacc" and "lex" progarms. These can help you develop some of the underlying lexing and parsing code and lets you worry about doing new stuff rather than redoing what's been done thousands of times already (the parsing and syntax checking). They were originally Unix programs but I think they've been ported to Windows and other OSs.

I would agree with what's already been said: If you really want to develop a compiler yourself because you think you have a market to sell it or just want to learn, go for it. If you're doing it to try to avoid buying one of the existing 8052 compilers it isn't worth it. The time you spend developing it is going to cost alot more than even the $2500 compilers that are out there.

Craig Steiner