RE: Junger et al.

I think we are in violent agreement here ...

Except that it seems odd that speech is no longer 1st amendment
protected as soon as it can be interpreted by a machine to do something.
So, then, if I want to deny you first amendment protections for
something, I can simply write a compiler to turn your words into machine
executable code, and suddenly, your words are no longer protected
speech.  How could that be a reasonable interpretation of functional
versus not?

Secondly, Gwin said that encryption is a special class of software which
is MORE functional.  This is definitely a misunderstanding, to say the
least.  I don't see how any particular class of software is necessarily
more or less functional than other classes of software.  In the
functional sense, all software, when compiled and executed is
functional, period (whether it performs according to its original design
is irrelevant).

Ern

	-----Original Message-----
	From:	Ed Gerck [SMTP:egerck@laser.cps.softex.br]
	Sent:	Monday, July 06, 1998 9:09 PM
	To:	Ernest Hua
	Cc:	cypherpunks@toad.com; 'cryptography@c2.net'
	Subject:	RE: Junger et al.

	On Mon, 6 Jul 1998, Ernest Hua wrote:

	>So today, I can write the following:
	>
	>1.	Find a container.
	>2.	Fill container with explosive substance.
	>3.	Move container to target location.
	>4.	Detonate container.
	>
	>As soon as I have a compiler and a target machine that can
execute these
	>instructions, suddenly this is not speech.  However, before
this
	>compiler and machine combo exists, event the electronic form of
this is
	>speech!
	>
	>How could this be?

	Gwin has written a phrase which deserves more analysis, IMO --
free
	from political overtones if we want to be impartial. The phrase
can
	be reworded as:

	  "source code is a device, that actually does a function"

	The difference and importance here is between syntatic and
semantics. 

	Your 4-instruction source code above is not a device today -- it
	cannot perform any function. It has only syntatics, not the "how
to". 
	But, if there were a machine that could supply the proper
semantics
	(ie, actually perform the functions 1-4) then your source code
above
	would be a device.  Further, your source code may not be a
device
	today but be a device tomorrow.

	As another example, bringing together one pound of inert metal
with
	another one pound of the same inert metal was not considered to
be
	explosive -- until U235 was used for the inert metal and
properly
	compressed. The difference is semantic, not syntatic. 

	In that, Gwin is correct. Can the source code actually perform a
	function?  Then, it is a device. Irrespective of the needed
platform,
	in the same way that an electric shaver is a device irrespective
of
	the local availablity of an appropriate power outlet.

	IMO, even though I consider Gwin to be correct to a very large
	extent, widespread use of crypto will not come from lifiting
such
	bans ... but from real need -- which does not outweigh the
hassle,
	today. The EFF has a wrong target there. 

	Do you know how much Internet e-mail traffic is encrypted today?
Can
	you believe less than 10%? Notwithstanding the rethoric
exercises and
	limelight it may provide, talking about encryption export bans
may
	not be as effective as desigining better and easier uses of
strong
	crypto -- that can then really drive market, legislation and
courts. 

	Need to use is a better key than need to know, it seems.

	Cheers,

	Ed Gerck

	>
	>Ern
	>
	>	-----Original Message-----
	>	From:	jkthomson [SMTP:jkthomson@bigfoot.com]
	>	Sent:	Monday, July 06, 1998 7:12 PM
	>	To:	cypherpunks@toad.com
	>	Subject:	Junger et al.
	>
	>
	>
	>	Reuters
	>
	>	3:40pm  6.Jul.98.PDT WASHINGTON -- A district court has
	>dismissed a law
	>	professor's challenge to US regulations strictly
limiting the
	>export of
	>	computer data-scrambling technology.
	>
	>	Judge James Gwin ruled late Friday that the export
limits, which
	>prevented
	>	Case Western Reserve University Law School professor
Peter
	>Junger from
	>	posting the text of encryption programs on the Internet,
did not
	>violate
	>	the constitutional right to free speech.
	>
	>	The Ohio court's ruling contradicts a California
district court
	>ruling last
	>	August that said source code -- the instructions a
person writes
	>telling
	>	the computer what actions to perform -- constitutes a
form of
	>speech
	>	subject to First Amendment protection.
	>
	>	"Unlike instructions, a manual or a recipe, source code
actually
	>performs
	>	the function it describes," Gwin wrote. "While a recipe
provides
	>	instructions to a cook, source code is a device, like
embedded
	>circuitry in
	>	a telephone, that actually does the function of
encryption."
	>
	>	Neither ruling gave speech protection to compiled code,
a
	>version of source
	>	code converted into an actual software program that
could be run
	>on a
	>	computer.
	>
	>	The US government appealed the California decision and
the issue
	>may
	>	ultimately be decided by the Supreme Court.
	>
	>	Civil libertarians and high-tech companies had hoped the
court
	>would
	>	overturn the export limits on encryption technology,
which uses
	>	mathematical formulas to scramble information and render
it
	>unreadable
	>	without a password or software "key."
	>
	>	Once the realm of spies and generals, encryption has
become an
	>increasingly
	>	critical means of protecting electronic commerce and
global
	>communications
	>	over the Internet.
	>
	>	But law enforcement agencies, fearing encryption will be
used by
	>terrorists
	>	and international criminals to hide their activities,
have
	>instituted
	>	strict controls to limit the export of strong scrambling
	>products.
	>
	>	Lawyers opposed to the export rules said the Ohio court
	>misunderstood the
	>	difference between source code and compiled code.
	>
	>	"The Ohio court clearly doesn't understand the
communicative
	>nature of
	>	software," said Shari Steele, an attorney with the
Electronic
	>Frontier
	>	Foundation. "It's true that software helps to perform
functions,
	>but it
	>	does so by telling computers what to do.... It certainly
is
	>speech
	>	deserving of the highest levels of First Amendment
protection."
	>
	>
	>
	>

>-----------------------------------------------------------------------
	>	 james 'keith' thomson <jkthomson@bigfoot.com>
	>www.bigfoot.com/~ceildh
	>	 jkthomson:C181 991A 405C EAFB 2C46 79B5 B1DC DB78 8196
122D
	>[06.07.98]
	>	 ceildh   :1D79 59AF ED75 5945 6003 8240 DA34 ACCA 9DE4
6BC9
	>[05.14.98]
	>	 ICQ:746241 <keys> at pgp.mit.edu     ...and former
sysop of
	>tnbnog BBS
	>

>-----------------------------------------------------------------------
	>	Technology is introduced, utilized, depended upon,
obsolete, 
	>	standardized, and understood, in that order
	>

>=======================================================================
_
	>


______________________________________________________________________
	Dr.rer.nat. E. Gerck
egerck@novaware.cps.softex.br
	http://novaware.cps.softex.br
	    --- Meta-Certificate Group member, http://www.mcg.org.br ---

On Mon, 6 Jul 1998, Ernest Hua wrote:

>I think we are in violent agreement here ...
>
>Except that it seems odd that speech is no longer 1st amendment
>protected as soon as it can be interpreted by a machine to do something.
>So, then, if I want to deny you first amendment protections for
>something, I can simply write a compiler to turn your words into machine
>executable code, and suddenly, your words are no longer protected
>speech.  How could that be a reasonable interpretation of functional
>versus not?
>

By the same reasoning as in my posting:

>>
>>As another example, bringing together one pound of inert metal with
>>another one pound of the same inert metal was not considered to be
>>explosive -- until U235 was used for the inert metal and properly
>>compressed. The difference is semantic, not syntatic. 
>>

Thus, your instructions to bring together those two one-pound pieces
would be your right to free speech -- maybe it is poetry -- as long
as that is not recognized as a federal crime! After it boooms the
first time... it is a device, and it is immediately recognized to be
an unlawful one. 

>Secondly, Gwin said that encryption is a special class of software which
>is MORE functional.  This is definitely a misunderstanding, to say the  

I could not find that quote -- "more functional".

BTW, I think that Gwin was struggling to express the notion that
source code (syntax) is a device whenever there is a clear binding
between that source code and known semantics, together with proper
pragmatics (the enviroment, as defined in semiotics), in order to
perform the desired function. 

This is called a Just In-Iime Compiler ... and Gwin is 100% right
when he affirms that the difference between source code and compiled
code is null regarding its effects -- its desired function. 

Thus, if the device is unlawful then it must be unlawful both as a
source code and as a compiled code. 


Cheers,

Ed Gerck


______________________________________________________________________
Dr.rer.nat. E. Gerck                     egerck@novaware.cps.softex.br
http://novaware.cps.softex.br
    --- Meta-Certificate Group member, http://www.mcg.org.br ---

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, 6 Jul 1998, Ernest Hua wrote:

> I think we are in violent agreement here ...
> 
> Except that it seems odd that speech is no longer 1st amendment
> protected as soon as it can be interpreted by a machine to do something.
> So, then, if I want to deny you first amendment protections for
> something, I can simply write a compiler to turn your words into machine
> executable code, and suddenly, your words are no longer protected
> speech.  How could that be a reasonable interpretation of functional
> versus not?
> 

Obviously there is no difference between 10 lines of PERL and
1500 lines of assembler (or machine code).  Executability is in the
eye of the beholder.  I can create an interpreter to execute english
(and in fact have done so in LISP).  Does that mean english is no longer
protected?

By the same argument, DNA is also a device.  The federal government
can now determine who you can reproduce with. (eugenics?)

DNA strands are instructions, code if you will, that describe to
cellular automata (literally ;-), how to (1) create a human being
(that could be dangerous -- they've been known to kill) and
(2) through fantastic complexity generate a fully functioning
human being, with all the behavioral patterns and biological
operations necessary to sustain said human being.

Imputing "deviceness" to code is missing the point.  There is
"deviceness" in all information.  It just depends on the context.

I would argue that it is specifically this "deviceness" that is protected
in the first amendment.  Speech isn't just about yammering on about
abstract concepts.  The imperative form of speech is probably the
most necessary to protect in a free society.

The most sacred form of democratic speech is probably the
vote.  Does it contain "deviceness"?  Absolutely.  The vote
is democratic force.

Is it not protected speech?

Speech in action is code in action.  When they can make code
illegal, they can make comittees to un-elect illegal and they
can make talking about impeaching the president illegal.

Think about that.

PostScript: It seems its not so much speech they don't like.
They don't like us talking to our computers, because there are
some things that computers do that are very powerful.  Maybe
more powerful than the state can handle.

In other words, I can tell Jan Hammer in Germany exactly how
to do what PGP does.  It might take him an infinity to actually
do it, but that speech is protected.  When I tell Jan's computer
how to do it, its illegal.

So speech is ok as long as its ineffective.  If it can actually
effect change then its illegal.

Interesting notion of free speech.

jim


-----BEGIN PGP SIGNATURE-----
Version: PGP 5.0i
Charset: noconv

iQA/AwUBNaJpoulhVGT5JbsfEQJU/gCgziy0U50OJXycwYNZPd4XsUuQfP8An3NY
BiWWWH8zvcvdVgKf2AM9ANu6
=6abs
-----END PGP SIGNATURE-----

Jim Burnes writes:
> Obviously there is no difference between 10 lines of PERL and
> 1500 lines of assembler (or machine code).  

One of Peter Junger's examples (I think used in the case) is the RSA
in 2/3 lines.  He actually obtained the US export administrations
written decisions on which of a small collection of titchy programs
was exportable.  RSA in 3 lines of perl they stated was illegal to
export.

So fun things with RSA are possible because they have decided that it
is not exportable, so perhaps you could export it on a floppy, or as
your .signature, with media cameras rolling, and try to get yourself
arrested for willful violation of dumb export laws.

Adam
-- 
Have you exported RSA today? --> http://www.dcs.ex.ac.uk/~aba/rsa/

print pack"C*",split/\D+/,`echo "16iII*o\U@{$/=$z;[(pop,pop,unpack"H*",<>
)]}\EsMsKsN0[lN*1lK[d2%Sa2/d0<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<J]dsJxp"|dc`

Greetings,

	There are t-shirts out there with both human readable and
machine readable (bar code) versions of the RSA in 3 code....  Wear it
when boarding an international flight....  They even state that they
are a munition and ITAR controlled in BIG letters on the back...


	Chris


On Wed, Jul 08, 1998 at 11:04:44AM +0100, Adam Back wrote:
> 
> Jim Burnes writes:
> > Obviously there is no difference between 10 lines of PERL and
> > 1500 lines of assembler (or machine code).  
> 
> One of Peter Junger's examples (I think used in the case) is the RSA
> in 2/3 lines.  He actually obtained the US export administrations
> written decisions on which of a small collection of titchy programs
> was exportable.  RSA in 3 lines of perl they stated was illegal to
> export.
> 
> So fun things with RSA are possible because they have decided that it
> is not exportable, so perhaps you could export it on a floppy, or as
> your .signature, with media cameras rolling, and try to get yourself
> arrested for willful violation of dumb export laws.
> 
> Adam
> -- 
> Have you exported RSA today? --> http://www.dcs.ex.ac.uk/~aba/rsa/
> 
> print pack"C*",split/\D+/,`echo "16iII*o\U@{$/=$z;[(pop,pop,unpack"H*",<>
> )]}\EsMsKsN0[lN*1lK[d2%Sa2/d0<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<J]dsJxp"|dc`

On Wed, 8 Jul 1998, Chris Liljenstolpe wrote:

> Greetings,
> 
> 	There are t-shirts out there with both human readable and
> machine readable (bar code) versions of the RSA in 3 code.... 

	RSA in 3 or RSA in 5?  Is there a new version?

	Now, all *real* Nettites know that the original 
	shirts were by Joel Furr, are no longer for sale, 
	and have an X-ed out Bill of Rights (or excerpts
	thereof) on the back, rather than any of this ITAR
	stuff.  And!!  It glows in the dark.  Rock rock.
 
						-Caj

	"This shirt is a ... " [squints] "... munchkin?"