|[4 earlier articles]|
|Re: Compiler project needed email@example.com (2000-02-22)|
|Re: Compiler project needed firstname.lastname@example.org (2000-02-23)|
|Re: Compiler project needed email@example.com (Stephen Sulzer) (2000-02-23)|
|Re: Compiler project needed firstname.lastname@example.org (Juergen Kahrs) (2000-02-23)|
|Re: Compiler project needed email@example.com (Daniel C. Wang) (2000-02-27)|
|Re: Compiler project needed firstname.lastname@example.org (2000-02-27)|
|Re: Compiler project needed email@example.com (Srineet) (2000-02-27)|
|Re: Compiler project needed firstname.lastname@example.org (Franck Pissotte) (2000-02-28)|
|Re: Compiler project needed email@example.com (2000-03-06)|
|Re: Compiler project needed firstname.lastname@example.org (2000-03-06)|
|Re: Compiler project needed email@example.com (Peter Wilson) (2000-03-06)|
|Re: Compiler project needed firstname.lastname@example.org (2000-03-06)|
|Re: Compiler project needed email@example.com (2000-03-06)|
|Date:||27 Feb 2000 02:35:35 -0500|
> I'm studying computer science on a Danish university and I'm going to
> write a compiler as a project in a course I'm taking.
> BUT! I totally blank for ideas :-( I don't want to write yet another
> Java compiler, so are there anyone with a really good idea as to what
> compiler I should write?
I too am a student of computer science and am working on a project
to build an optimizer generator. More specifically, we plan to design
a tool that'll be helpful in building optmizers just as Yacc is useful
for building parsers. We have developed a minimal pure functional
specification language, which also supports dataflow analysis. If
anybody is interested, I'd really like to describe it. Also, we have
designed a nice (I think,) way to interface it to existing compiler
Well, enough 'bout my project.
If you are interested in doing something new, an interesting area
of study is how to use profiling information to do optimizations. That
is, you''ll be writing an environment that sees how the program runs
for various inputs, given by the user of course, use the profiling
information (space used, time used, which part of code runs most of
the time,) to perform optimizations on the program. I'd suggest trying
it for a simple interpreted language, so you don't spend much time
writing the parser. Gopher, though not very small, may be a good
candidate since the existing interpreter gives you some profiling
information. You might want to check it out.
A caution though, this project involves research, and if your
school focuses more on implementation and coding than the idea and the
research involved, think twice and consult as many experts as you
Hope this helps. Please keep us informed of what you choose.
Return to the
Search the comp.compilers archives again.