PACT in GSoC 2012

[Brian Charles Gernhardt]

Hello all,

I suppose not everybody remembers me from last summer, or from #parrot over the last year, so I'm going to re-introduce myself.

My name is Brian Gernhardt and I'm an MS Student at the Rochester Institute of Technology.  My current focus is programming language theory and compilers, and my thesis[1] is on compiler optimization.  Next year, I will be in a PhD program (school TBD) to do additional research in this field.

Last summer, I worked on porting PAST to NQP so it could use bacek's work on directly generating bytecode[2].  The results of this were somewhat mixed[3].  At the end of that project, and sporadically since then, I've worked on designing a system I've called the Parrot Alternate Computer Toolkit (PACT)[4].  It's a rethink of our currently HLL compiler libraries, designed from the ground-up to directly output bytecode instead of relying on the aging IMCC.  This is an ambitious project, but I think it could bear fruit rather quickly.  The lowest levels could be used to create a PIR alternative, a disassembler, and possibly a backend for a skilled compiler writer.

I hope the community will support me using Summer of Code to start this project.  In one summer, I'm fairly certain I could create a new assembly language for Parrot that has a useful set of classes that could be used in other project and future layers of PACT.  Of course, I'll write up a formal proposal during the application process but wanted to start floating the idea now.

[1] http://www.cs.rit.edu/~bcg2784/
[2] https://github.com/parrot/parrot/tree/nqp_pct
[3] http://www.parrot.org/content/end-gsoc-not-time-stop...
[4] https://github.com/parrot/PACT

~~ Brian Gernhardt
benabik on #parrot, #perl6
MS Student, RIT CS