[antlr-interest] related master's thesis ideas...

[Terence Parr]

On Mar 2, 2006, at 9:21 PM, Kaleb Pederson wrote:
> 1) Incremental parsing (specifically for not re-parsing an entire  
> file, for
> example within an IDE).  This could be interesting, both in theory  
> and to try
> to make Antlr support this.

Prashant  and I are going to be  publishing a paper on this soon so I  
would work on something else. ;)

> 2) Create a static code analysis tool, probably to handle C/C++  
> code although
> if I used the right IR trees, it could probably easily be expanded  
> to other
> languages fairly easily.  This would probably also present the  
> ability to
> create a mini-language of sorts, something that would allow you to  
> create
> specific test cases that the programmer was interested in, with a  
> certain
> standard set of featurs.
>
> 3) Something with Antlr-v3 and probably/hopefully Python ;),  
> perhaps wrapping
> the C/C++ generated code so that it can be used as an extension  
> module,
> assuming they were considerably faster than the regular Python one  
> (only real
> reason to do such)?

Perhaps evaluating rewrite systems for v3 like

"int <ID> = <expr>;" -> "int <ID>;\n<ID> = <expr>;"

> PS: Ter, I hope your dvorak skills are coming along.  I switched to  
> dvorak
> about two and a half years ago (maybe three now) and I'm probably  
> 20-40%
> faster than I was back then, and I was quite capable then.  It's  
> definitely
> worth it (although the efficiency takes a while to pick back up)!   
> Keep at it
> and good luck!

  thanks for the encouragement.  I memorized the keyboard and one  
point five hours but am having difficulty getting my efficiency up.  
Are you able to use emacs with the new keyboard too?

Ter