[2073] in java-interest
Re: java-interest-digest V1 #179
daemon@ATHENA.MIT.EDU (Andy Johnson)
Fri Sep 22 18:23:13 1995
To: jmarsh@mitre.org (John Marsh)
cc: java-interest@java.sun.com
In-reply-to: Your message of "Thu, 21 Sep 1995 13:35:48 PDT."
<199509212035.NAA15449@java.sun.com>
Date: Fri, 22 Sep 1995 15:14:20 -0400
From: Andy Johnson <andyj@osf.org>
>
> From: jmarsh@mitre.org (John Marsh)
> Date: Thu, 21 Sep 1995 13:35:10 -0400
> Subject: Re: Decompilation of JAVA bytecode to source
>
> Java'ers
>
> Unfortunately I lost the thread that this email really responds to...
>
> As previously noted, it is possible, though not particularly easy, to
> decompile bytecode to source.
>
> My understanding is that when the architecture neutral distribution format
> (ANDF) people at OSF developed the ANDF spec, they spent a lot of time
> ensuring that it was "not possible" to decompile ANDF code to source. They
> were extremely concerned about intellectual property issues associated with
> giving people mechanisms that permitted access to source. They believed
> that if it were not possible to protect source, it would be significantly
> more difficult to get vendors to sign up to deliver their code in ANDF. (Of
> course ANDF never went anywhere, but that is/may be a different issue...)
>
> It seems to me that it would not be a good idea to ignor the ANDF work on
> source access as JAVA goes forward, otherwise it might limit the potential
> for the future.
> John Marsh
> The MITRE Corporation
>
> jmarsh@mitre.org
> Phone (703)883-7086
> FAX (703)883-3315
>
>
I would suggest that you check out http://www.osf.org/andf for more information about ANDF. Of particular interest on this subject is the paper "Protecting Source Code with ANDF" by Stavros Macrakis. The paper is available in:
ftp://riftp.osf.org/pub/andf/andf_coll_papers/ProtectingSourceCode.ps
The contention is that Binary protects source code but is architecture dependent. ANDF protects source equally well, but is architecture independent. In essense, there is some information in the ANDF that is not directly contained in an executable binary, but can be inferred (such as data types). The best that one can hope for would be an "obfuscated" (or "shrouded") C program, since no identifiers are present except for external names.
In viewing the de-compiled Java byte code, some identifiers are present (notably class members), which would enhance the quality of a de-compiled program somewhat, but it still wouldn't be as readable (or maintainable) as the original source program.
-AndyJ
E. Andrew Johnson email:andyj@osf.org
OSF Research Institute
11 Cambridge Center Tel: (617) 621-7378
Cambridge, MA 02142 Fax: (617) 621-8696
-
Note to Sun employees: this is an EXTERNAL mailing list!
Info: send 'help' to java-interest-request@java.sun.com
