RE: What is the future of Compiler ?

Quinn Tyler Jackson <>
20 Jun 2006 10:30:32 -0400

          From comp.compilers

Related articles
What is the future of Compiler ? (Inderaj Bains) (2006-06-15)
RE: What is the future of Compiler ? (Quinn Tyler Jackson) (2006-06-20)
| List of all articles for this month |

From: Quinn Tyler Jackson <>
Newsgroups: comp.compilers
Date: 20 Jun 2006 10:30:32 -0400
Organization: Compilers Central
References: 06-06-049
Keywords: design
Posted-Date: 20 Jun 2006 10:30:32 EDT

The question:

Where the compilers are headed? This is a very broad topic, I will try
to address some parts

First, most problems in compilers are intractable (NP), this ensures
heuristics will always be needed and developers and researchers will
have jobs.

In terms of research, Scanning and parsing have had a lot of research
and are thus fairly matured. Optimizations is going to be an active
research area forever with newer things regularly coming up like using
machine learning for Heuristics.

A highly biased response:

I believe that a great part of the future of compiler technology is in
adaptive parsing. I am biased because AP is my main CS love. I'm crazy
enough to believe that separation of syntax and semantics is a Bad Thing if
we consider the kinds of parsing problems that fall just outside of
traditional compiler tasks. (Parsing of RNA pknots for example. Parsing of
funky scripting languages like Perl without using perl to do it, for

Grammars (and parsers) that learn, grow, adapt, et cetera, in my thinking,
are "a future" of compilers, though not "the" future.

For more on this, Adapting to Babel is now available for online reading at:

or directly at

(Here's to hoping that URL doesn't wrap.)

It covers RNA and Perl, as well as a lot of other difficult to parse
languages (without requiring ad hockery in "the code" -- one of my Holy

The online reader requires JavaScript to be enabled for the
* domain. The bookmarker works only if cookies are
enabled for that domain.

I apologize if some of the text is a bit hard to read -- zooming to 125%
might help in some cases. It's the best I could manage without making the
image files way too large for the server. Future versions of the dropdown
index will include all chapters, rather than just the ones that are there

The Meta-s system is currently being integrated into industrial strength
systems -- I can't reveal where right now due to NDA -- but it is starting
to find some legs in industry. (This isn't a promotion -- I'm just
explaining that it's not entirely a lab-only thing at this stage.) It's
holding its own against some pretty tough contenders. (Someone in this list
recently asked about parser generators with IDEs -- this is one -- but we
don't currently offer the IDE for sale.)

Anyway -- that's my biased view. Others' mileage may vary.

(Since the previewer is in beta -- please let me know at if you encounter anything egregious. Much


Quinn Tyler Jackson

Post a followup to this message

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