[Coco] FPGA 63x09

Jeff Teunissen deek at d2dc.net
Sat Mar 28 19:22:14 EDT 2009


Bill Barnes wrote:
> If what you say is true, well, EVERYONE is in trouble.

Not so much. Even when cloning a program, it's rare that there's an actual
copyright infringement...because you don't usually have access to the source
code and disassembling the original (which may preserve the program's
non-functional characteristics) is rarely done these days.

Creating a new product based on how another product behaves is not likely to
be an infringing act. Disassembling that product and reassembling it
(regardless of modifications you made) is virtually guaranteed to be.

If you look at a disassembly and figure out merely what the code does ("given
char pointers on the U stack, call I$WriteLn") and then do that without
consulting the original source, you're somewhat less likely to get screwed. If
you figure out what the code does and then someone _else_ implements it
without having even seen the original source or the disassembly, then you're
even safer...but that doesn't prevent you from being sued, it just gives you a
decent defense when it does happen.

In the case of the Green Mountain stuff, it actually looks like a booby trap
and/or a way to prevent competition from even showing up.

The license.txt included in the VHDL source says, flat out, that you can embed
the code into hardware you make, but you aren't allowed to distribute any part
of the source code to anyone, whether by electronic or any other means.
Presumably, they will allow you to sub-license the code IF you pay them for a
separate license.

Later on they say you can modify the source code, but that the modifications
are still subject to the terms of the license, i.e. you still can't give them
to anyone.

The relevant sections:

LICENSE TO USE SOURCE CODE.  Licensor  grants to Licensee the right to
use one copy of the Source Code of the  enclosed design.  Licensee may
not  copy and/or distribute the  Source Code or  any written materials
that accompany the Source Code.   Licensee may not provide nor utilize
any electronic access to provide the Source Code to any other entity.

MODIFICATION. The Source Code may  be modified by Licensee without
the prior written permission  of Licensor.  Any  modifications made to the
Software will continue  to be subject to  the terms and  conditions of
this License.

[snip]

>> Sorry, the point is not moot. You can change EVERY single line and still
>> infringe copyright, _especially_ with such "template" use.
>> Copyright does not apply at all to the so-called "functional aspects" of a
>> program -- that is, the stuff that is required to perform a task. If any
>> particular job can be performed optimally only by a specific method, no code
>> that exactly implements that method is, or can be (at least in the US),
>> covered by copyright.
> 
> Good thing too, as Look and Feel could get even M$ in trouble. However, many IP 
> lawsuits are over *how* code does a task.

That's what I said in the next paragraph. There's almost always more than one
irreducible way to perform a task in a high-level language, and those
different methods can be considered "speech".

If I like to write

if ((foo = bar (args))) {
}

and that generates slightly different binary code to

foo=bar(args);
if(foo)
{
}

then that slight difference in the _binary_ is an expression of my speech
(while all of the differences are expressions of speech in the source code).

Likewise, if I like to early-exit functions while someone else likes to use an
error-processing routine, that too is an expression of speech. All this is
(protected) "how" and "why" stuff, not just (unprotected) "what" stuff.

[snip]

>> By starting with someone else's design (even if you don't actually copy a damn
>> thing directly!), you open the door to copyright infringement, because the
>> meaning of the original code doesn't change when you add new code. Even if you
>> were to incrementally rewrite until nothing of the original code is left,
>> unless you completely change the design as well, it's still a derivative
>> work...and if you were going to do that, what was the point in using the
>> original code in the first place?
> 
> Sounds like SCO's argument against Linux, as some who contributed to Linux used to work on the code for Unix. But there has to be enough *original,* or code deemed the same, present to stick.

Caldera never really had an argument against Linux, especially considering
they never owned Unix to begin with.

Having access to Unix source code, then writing significantly similar "Linux"
code, is probably enough to be considered a derivative work (and thus
infringement, if you are not allowed to create such derivative works).

Unfortunately for Caldera ("SCO"), this only happened once (by an SGI
employee), and never hit a single running kernel. There were other similar
bits of code, but those came from other places...and in some cases, it was
AT&T that stole the code from BSD.

>> Some years back, George Harrison was sued for copyright infringement over a
>> song, "My Sweet Lord", that he had written on his own. Harrison was found by a

[snip]

> In a case like this, all one has to do is convince a jury you are right, and in
> reality, what you convince them of can be disguised BS just as easily as the 
> truth. Sad but true, and we all suffer for it.

Maybe, but it doesn't happen as much as you might think it does. Juries are
made mostly of pretty rational people. Besides, it's pretty clear that
Harrison *did* accidentally re-implement the melody.

[snip]



More information about the Coco mailing list