Re: xcdroast does no longer work with wodim: Who to blame?
On Thu, 5 Mar 2009, Russell Coker wrote:
On Thu, 5 Mar 2009, Bill Unruh <unruh@physics.ubc.ca> wrote:
Sorry, but that means that the concept does exist. The binary of a program
is a totally different animal from the original. Not a shred of similarity
in expression exists between the two. If I grep phrases from the original
in the binary, I will not find them.
In future please do some tests before making wild claims (or even just
consider how a compiler works).
Oh dear, Yes, I did think of that and assumed that the readers were
intelligent enough not to need obvious caveates. My programs have no string
literals. Yes, string
litereals are sometimes preserved. And the point is? YOu think that those
string literals will convert a non-copyrighted binary to a copyrighted one
(Read "Fair use")? "If the program has theree letters in common with the
source then the program is derivative and if two it is not?"
In all the common compilers every text string (IE something that is enclosed
in "" in the source) will appear exactly the same in the binary. Variations
on the compressed executable theme (which was common in MS-DOS days) will of
course be exceptions to this rule.
Who cares in this context?
$ gcc test.c
$ strings a.out
/lib/ld-linux.so.2
__gmon_start__
libc.so.6
_IO_stdin_used
puts
__libc_start_main
GLIBC_2.0
PTRh
[^_]
hello world
$ ./a.out
hello world
$ cat test.c
#include <stdio.h>
int main() { printf("hello world\n"); return 0; }
$
That program is probably not copyrightable. It has long gone into the public domain.
A binary is a translation of the source. Does it fall under the copyright on
the source or does it not? That is the issue of "derivative work". How close
does the "translation" or the use of the original have to be to still fall
under the original copyright?
And how different does it have to be not to do so.
Reply to: