{"id":828,"date":"2010-01-02T09:47:28","date_gmt":"2010-01-02T17:47:28","guid":{"rendered":"http:\/\/boost-spirit.com\/home\/?p=828"},"modified":"2010-01-03T09:52:07","modified_gmt":"2010-01-03T17:52:07","slug":"build-a-compiler-what-to-build","status":"publish","type":"post","link":"http:\/\/boost-spirit.com\/home\/2010\/01\/02\/build-a-compiler-what-to-build\/","title":{"rendered":"Build a Compiler, What to Build?"},"content":{"rendered":"

Thank you all for your warm feedback on the \u201cBuild a Compiler\u201d<\/a> post. It seems this has become very popular indeed. I guess it\u2019s time to start. Your overwhelming feedback and comments is enough motivation to carry on with the article series.<\/p>\n

In general, an imperative OO language seems to be the way to go. It\u2019s not surprising that C++ is very popular. People want a C++ parser! Barring that, due to complexity, a subset or a sanitized\/re-syntaxed C++ (e.g. SPECS) is also a popular request. Go is also quite popular. That language indeed looks good and modern. FP, especially LISP\/Scheme and even Haskell(!) is also quite popular. And hey: Javascript! and Python! Life would not be complete without these fun languages:-).
\n
\nHere\u2019s a tally of languages suggested for the project sorted by popularity:<\/p>\n

I\u2019ll need some time to think about all these. In the meantime, it\u2019s still not too late to chime in with your comments. This particular post will be dynamic. I will add on it over time based on my thoughts and your feedback.<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
C++<\/td>\n4<\/td>\n<\/tr>\n
Lisp\/Scheme<\/td>\n4<\/td>\n<\/tr>\n
C<\/td>\n3<\/td>\n<\/tr>\n
Go<\/td>\n3<\/td>\n<\/tr>\n
Haskell<\/td>\n3<\/td>\n<\/tr>\n
C++ SPECS<\/td>\n2<\/td>\n<\/tr>\n
Javascript<\/td>\n2<\/td>\n<\/tr>\n
Python<\/td>\n2<\/td>\n<\/tr>\n
ALGOL-like<\/td>\n1<\/td>\n<\/tr>\n
C++ Subset<\/td>\n1<\/td>\n<\/tr>\n
Chapel<\/td>\n1<\/td>\n<\/tr>\n
Clojure<\/td>\n1<\/td>\n<\/tr>\n
EBNF<\/td>\n1<\/td>\n<\/tr>\n
GIMPLE<\/td>\n1<\/td>\n<\/tr>\n
Forth<\/td>\n1<\/td>\n<\/tr>\n
Java<\/td>\n1<\/td>\n<\/tr>\n
Kaleidoscope<\/td>\n1<\/td>\n<\/tr>\n
Lua<\/td>\n1<\/td>\n<\/tr>\n
OCC<\/td>\n1<\/td>\n<\/tr>\n
Pascal<\/td>\n1<\/td>\n<\/tr>\n
Perl<\/td>\n1<\/td>\n<\/tr>\n
R<\/td>\n1<\/td>\n<\/tr>\n
SML<\/td>\n1<\/td>\n<\/tr>\n
Scala<\/td>\n1<\/td>\n<\/tr>\n
TinyC<\/td>\n1<\/td>\n<\/tr>\n
VB<\/td>\n1<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
Requirements:<\/h6>\n

Here are the requirements thus far:<\/p>\n

    \n
  1. It will be a statically typed, imperative language of the ALGOL\/Pascal\/C++ family.<\/li>\n
  2. It shall be embeddable in your application.<\/li>\n
  3. Seamless integration with C++ will be provided with full round-trip calls to and from C++.<\/li>\n
  4. There will be zero runtime overhead and zero requirements\/dependencies apart from the C++ standard lib and the LLVM back-end.<\/li>\n
  5. It shall target LLVM with its JITC support.<\/li>\n
  6. It shall be a practical tool, not a toy for instructional purposes only.<\/li>\n<\/ol>\n

    Share this:<\/h3>