Re: extensible compilers

In 98-03-155 mtrofimov@glas.apc.org wrotes:
>Do you know anyhow about extendsiable compilers? In the book
>"Algorithms+[DS...]" Wirth wrote that is _VERY_ impossible (not quite,
>but "very"). This time, I develope a compiler, which uses structured
>diagrams for syntax analyzing...
>
>Do you know an example of similar approach?


  I have developed and just started distribution of
"Extensible Java pre-processor", EPP. (http://www.etl.go.jp/~epp)


EPP has a recursive descent parser written with mixin-based
programming style. "EPP plug-ins" can extend Java syntax in
differential-way, using a kind of inheritance mechanism. Left
associative operators can be also added. (Sorry, currently there are
no technical document about the EPP parser, however, source code is
included in the distribution package.)




>[Extensible languages were fashionable in the mid 1970s. It's not
>that hard to do, but people found that the ability to write each
>program in a slightly different, incompatible, language wasn't all that
>useful. -John]


I am conscious of the incompatibility problem. The extension of
syntax should be as small as possible. I designed EPP to be a good
tool for programming language researchers. Syntax extension should be
done only by designers of good design sense. Good syntax will
increase productivity of softwares, and then, EPP must be useful tool.
---
Yuuji Ichisugi
  Computer Science Division
  Electrotechnical Laboratory
Email: ichisugi@etl.go.jp
http://www.etl.go.jp/~epp
--

Return to the comp.compilers page.
Search the comp.compilers archives again.