On Fri, Mar 11, 2005 at 03:32:12PM +0100, Michael Below wrote: > Humberto Massa <[EMAIL PROTECTED]> writes: > > > Despite the letter of the GPL and its post-amble, "linking", generally > > construed as "stitching together (normally executable) object (as > > opposed to source) files and resolving fixups so the result is an > > executable file" does NOT make a derivative work. Derivative works are > > made when you have intelligent *transformation* of the original > > work. Linking is not intelligent -- much au contraire, it's fully > > automatic. > > Hm. So the LGPL is completely useless in practice?
No, i think there are two different notions : 1) the fact of linking code together in a single binary file. You could obtain a similar effect with a plain : echo $SIZE_OF_A $SIZE_OF_B > archive cat a >>archive cat b >>archive and not make a and b derivative works of each other. 2) the fact of A making use of the functions provided by library B, that they are linked together dynamically or statically. This is the notion that the LGPL addresses, which is totally orthogonal to the issue 1), which is mere aggregation in a commom media volum or medium of distribution. > > So, no, if it doesn't fit, you must acquit -- IOW: the fact of > > embedding the flasher and the flash in the same ELF file does not make > > the combined work a derivative work on any of them; only a > > "collective" work on both. > > I think you are right, we are talking about a collective work. But I > still believe that the GPL demands the distribution of the flash image > under GPL terms, when both image and flasher are distributed together. Nope, since it clearly exludes it in the last paragraph of clause 2 : In addition, mere aggregation of another work not based on the Program with the Program (or with a work based on the Program) on a volume of a storage or distribution medium does not bring the other work under the scope of this License. So, there can be no doubt. Friendly, Sven Luther -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]