Re: GPL and "free" software

[Ed Avis <epa98@doc.ic.ac.uk>]

Joshua E. Rodd wrote:

>>But most of the time, header files don't actually include code.
>
>FD_SET certainly includes code.

Okay.  I did say, 'most of the time'.  For the sake of argument, let's
just discuss 'pure' header files that don't contain actual code, (or
maybe the code is so trivial and short as to be not copyrightable).
 
>>They
>>include function prototypes, which aren't strictly necessary (K&C
>>managed for years without them), declarations of functions (which are
>>just there to reassure the compiler that 'printf()' isn't a typo), and
>>constant values such as #define SIGINT 2, which being just a few
>>simple numbers, probably aren't copyrightable by themselves.
>
>Huh? One could as well argue that an object file is `just a bunch
>of unsigned numbers'. It's still copyrighted.

An object file is several thousand rather long integers.  I meant that
something like:

#define BUFFER_SIZE 1000
#define MAX_THINGIES 6

probably isn't copyrightable, especially once you strip away the names
and are just left with the pair of integers '1000, 6'.  You could make
an enormous header file with #define FIRST_INSTRUCTION 0x28a45e92,
#define SECOND_INSTRUCTION 0xf3485dea, and so on, which would be big
enough to copyright.  Of course, IANAL.

>>Prototypes and declarations don't end up in the final object code -
>>they're just 'scaffolding' for the compiler.
>
>They very well could end up in the final code--think about Java.

Java doesn't have separate prototypes or header files.  AFAIK the
information on a method's name, what arguments it takes, and so on,
are stored together with the bytecode for that method - but then, they
came from the same source file anyway.  It's not a case of my object
code getting 'polluted' with copyrighted material from somebody else's
header files.

-- 
Ed Avis
Advertise here! epa98@doc.ic.ac.uk