TODO

http://pdos.csail.mit.edu/~baford/packrat/
http://www.codeproject.com/KB/recipes/grammar_support_1.aspx
http://www.replicator.org/node/67
http://www.codeplex.com/irony

http://www.eqatec.com/tools/profiler
http://code.google.com/p/nprof/
ms (but team version)
http://www.wikihow.com/Optimize-Your-Program%27s-Performance
http://stackoverflow.com/questions/375913/what-can-i-use-to-profile-c-code-in-linux/378024#378024

http://www.thinkingparallel.com/2007/04/18/ten-questions-with-joe-duffy-about-parallel-programming-and-net-threads/


* Try reworking ftests:
	- make sure that the various options work
	- make sure clean gets rid of the generated files
	- can we (for testing at least) specify both c# and f# in the peg files?
		'f#'`blah` 'c#'`stuff`
* Work on better language support:
	- need a language setting
	- some setting are language specific
	- add a warning for ignored setting
	- may want a --silent option
	- update the grammar in the template readme
	- add support for c++ (use member function pointers)
	- add support for f#
* May want to talk about SS.
* Might want some tips & techniques ftests.
* Might be neat to allow peg-sharp to create dot files (using graphviz)
* May want to rewrite the readme using asciidoc.
* Do a release.
* Update the wiki (eg c++ and f# support)

* Maybe also rtf generation
* Look into optionally generating c++ code.
* Might want to consolidate ftests into a single executable.
* Do a release.

* Should there be an explicit way to match eot?
* Might be nice to add some support for error recovery so multiple parse errors can be
   reported. See Bison's simple error recovery.
* Would be nice to support #if stuff.
* May want to support utf-16 (could look for zeros or a bom).
* Might want a --setting switch. Should override earlier switches and peg file settings.
* Allow arbitrary sequences (instead of just strings) to be parsed?
* Would be nice if we could do optimizations like rule inlining but this is tricky with the
   way we handle semantic actions (they are attached to rules not expressions, rely on
   the order of expressions within the rule, and default to returning the value of the
   first subrule).
* Release checklist:
	- make sure the csproj has the right version
	- make sure sln still builds