Wow and I though it was an innocent question.

Here is my view

1. Obfuscation doesn't prevent ppl from stealing source.  Java has the same
problem, there are plenty of great decompilers for java to prove it.
2. Though we understand point 1, "the perception" is that perl being a
scripting language creates a "risk" and large companies are risk-averse,
which is why large companies prefer java regardless of the lack of security
it provides(larry wall doesn't have the same marketing budget as SUN).

So if we accept point 1 & 2 we come up with a few interesting questions.

1.  If someone is without clue enough to use B::Deparse or B::Deobfuscate,
are they also without clue enough to understand that par extracts everything
into a temp directory?
2.  Will using B::Bytecode and B::ByteLoader provide some level of
"obfuscation" above extracting to TMP, and also provide a performance boost?
3.  Will using some form of real encryption(DES,3DES,etc.) to the files
extracted in TMP provide any more security (beyond perceived) than
obfuscation?  If you recompile perl58.dll you can do what ever you like with
what's in memory, and if you can manipulate memory you have the unencrypted
source.
4.  Will Perl6+Parrot make PAR obsolete.
5.  If 4 is true, how long will it be before perl6 is done (current
estimates are 2005-2008), and is anyone willing to wait.

>From the little poking around I've done it seems that B::Bytecode is the
best direction to go.  The work being done with Parrot and Perl6 has very
Dramatic performance increases, though it is yet to be determined if
Parrot+Perl6 will alleviate the need for PAR.

The problem with B::Bytecode for me is that it seems to be very broken on
win32 and 5.8.0, and until activestate releases a 5.8.1 build, it will be
impractical to do any testing on win32.  Until I get 5.8.1 to build on my
system I will not be able to do any testing, though it may be worth it for
me to test on linux 5.8.1 just to get a feel for what will be required to
deal with B::Bytecode and B::ByteLoader, I can't imagine that it will be
that hard to introduce a step that takes $pfile and post processes it with
B::Bytecode.

-jess

-----Original Message-----
From: Tommy Butler [mailto:[EMAIL PROTECTED]
Sent: Saturday, October 18, 2003 4:07 PM
To: [EMAIL PROTECTED]
Subject: RE: Source Hiding (was: Encyrpting Source)



--- "Mr. Lindsay Morris" <[EMAIL PROTECTED]> wrote:
> We're getting pretty far off-topic for a PAR list.

I think we're very much on topic here.  PAR by nature is such a commodity to
the great majority of those who use it _because_ they can distribute their
code
(via PAR) in a manner that isn't as open as pure source files.  This
discussion
is simply touching on an inherent theme (with respect to PAR) that
definitely
deserves some public discussion.  I am excited to have heard from Randall,
and
the opinions of other Perl hackers are likewise valued.  I want to hear what
the community has to say about this.

Randall (and those of his mindset regarding this issue), would you also
avoid
purchasing a compiled Perl 6 executable?  I believe they can be easily
decompiled however using some features from O:: or B:: maybe, though that
isn't
the point.

Friends, please share your views on this matter.  This is not flamebait.
This
is an issue that deserves some attention and I for one want to see what
others
have to say about this topic and better understand why and where you are
using
PAR.

Thanks.

=====
Tommy Butler <[EMAIL PROTECTED]>
phone: (817)-468-7716
6711 Forest Park Drive
Arlington, TX  76001

Reply via email to