Re: GPL traitor !

[Andrew Halliwell]

David Kastrup <dak@gnu.org> wrote:
> Hyman Rosen <hyrosen@mail.com> writes:
> 
>> David Kastrup wrote:
>>> An executable image stored in parts is still an executable image.  Just
>>> because the assembly happens on-demand does not change the intent.  As
>>> long as there is no conceivable use without the (automatic and planned)
>>> assembly, the exact time frame of the assembly is not really relevant.
>>
>> You keep talking about intent, but even if that mattered,
>> which it does not, US copyright law specifically declares
>> the copying needed to execute a program on a machine to be
>> non-infringing.
> 
> But we are not talking about copying, but assembling.  The act of
> creating a mere in-memory copy does not cause additional worries, as
> this is the _intended_ use of the copy.  But dynamic linking is not mere
> copying, it is _assembling_ the separate parts into a coherent whole
> executable in a single memory space.

It isn't assembling anything. Assembly is the... oh what's the point, you
didn't respond to my last post.

It is LINKING. The library exists in memory. The program exists in memory.
All the linker does is allow the program to CALL the bits of the code it
needs FROM the library. 

The library itself is untouched by this linking process. Other programs in
memory may also link to exactly the same bits of code. 

That is the whole POINT of dynamic linking. To remove the need to have 40
programs all loaded into memory containing large chunks of identical code.
 
>> Further, you speak of "an executable stored in parts". But the
>> fundamental aspect that you erroneously choose to disregard is
>> that some of those parts, namely the GPL-copyrighted dynamic
>> libraries, are not being copied by the author of the program,
>> nor are they being distributed along with the program.
> 
> They are assembled with the other parts of the program at runtime,
> according to preordained instructions by the program author, as an
> unavoidable part of starting the program.

yes, but at NO POINT in the compilation or distribution process is the code
actually distributed WITH the program!

OK...

Look at it this way...
You're reading a text book. It makes references to other textbooks
throughout the text, at the back of the book, is a bibliography, pointing to
each of the books referenced. So, you collect ALL the books together, and
when a reference comes up, you put down the book you're reading, read the
passage in the other book that's been referenced, and then return to the
original book...

THAT is the author/book/reader equivalent of dynamic linking.

For static linking, each passage referenced would have to be copied and
printed verbatim from the source books into the book your reading. For short
passages that might be fine, but if the book referenced an entire chapter
here, and entire chapter there...

1: the book would be VERY big and bloated.
2: The book would be in breach of copyright of every author that was
referenced beyond a short passage.


 
> And on a demand-paged operating system (like pretty much every system
> nowadays), executables are loaded into the physical address space
> piecemeal and on-demand.  Whether or not you have used static or dynamic
> linking.

Irrelevant to the argument of copyright.

> Copying, not linking.  Linking is not just copying.
 
No copying need be involved in linking.
Therefore copyright is irrelevant in dynamic llnking. 
Only static linking falls under copyright law.
-- 
|                          |What to do if you find yourself stuck in a crack|
|  spike1@freenet.co.uk    |in the ground beneath a giant boulder, which you|
|                          |can't move, with no hope of rescue.             |
|  Andrew Halliwell BSc    |Consider how lucky you are that life has been   |
|           in             |good to you so far...                           |
|    Computer Science      |   -The BOOK, Hitch-hiker's guide to the galaxy.|