Lambda the Ultimate

inactiveTopic The PILER Decompilation System
started 7/12/2003; 11:29:43 AM - last post 7/13/2003; 2:21:55 AM
Ehud Lamm - The PILER Decompilation System  blueArrow
7/12/2003; 11:29:43 AM (reads: 1525, responses: 1)
The PILER Decompilation System
The PILER decompilation system is documented in P. Barbe, The Piler system of computer program translation, Technical report PLR-020, Probe Consultants Inc., September 1974. However, this document is probably only available in Microfiche form, and would be difficult for many to access. As the first attempt at a general decompiler, I thought it was worth summarising in a little more detail.

Posted to history by Ehud Lamm on 7/12/03; 11:30:45 AM

Andris Birkmanis - Re: The PILER Decompilation System  blueArrow
7/13/2003; 2:21:55 AM (reads: 360, responses: 0)
Meta-model of a machine code is so close to that of COBOL... One has to wonder, why should it take so much effort. Disclaimer: I know, there are weird machine codes, but somehow the article makes me believe the PILER worked for mainstream von Neumann architectures.

From the other side, imagine Java program. From the PILER's point of view, Java classes are data for JVM. For programmer, they are code. Is it just psychological difference? Is data fundamentally differrent from code? If we analize functional-language executable, what will be a COBOL source code, if not just incomprehensible transcript of machine code? Could the PILER exctract the original intent of the program? I doubt.

Yes, I understand it was just a tool for porting programs, but I always tend to over-theorize and put my two cents everywhere :)