On Tue, Feb 08, 2000 at 09:14:55PM -0500, Andreas Pour wrote: > Right, but for the analysis to be complete you must include the > definition of what the complete source code is. This is provided in > the second sentence of the ultimate para. in Section 3, which provides > > For an executable work, complete source code means all the source > code for all modules it *contains*, plus any associated interface > definition files, plus the scripts used to control compilation and > installation of the executable.
That's part of the definition -- it's not reasonable to claim that this somehow excludes any other definitional material from the GPL (or, for that matter, any other definitional material from common english usage). > The key part being the reference to "all modules it *contains*", > rather than all modules which may at run-time be linked to it. This is a distinction you've introduced, not something that's ever stated in the GPL. > To substantiate the point, I again refer to my Webster's New Universal > Unabridged Dictionary (2d ed. 1983) and look-up "contain", quoting the > relevant definitions: > > (1) to have in it; hold; enclose or include. 2) to have the > (capacity for holding. 3) to be equal or equivalent to; as, a > (gallon *contains* four quarts; > > Now, please explain how the "executable work" which I am distributing > (kghostview which is dynamically linked to Qt) "has in it", "holds", > "encloses or includes", "has the capacity for holding", or is "equal > or equivalent to", the Qt library. I don't know which copy of kghostscript you're distributing, but let me ask you this: do you expect that copy to work? > Sure, the Qt library can later be added to it -- like sugar to the > gallon of water -- but is not contained in it. We're not dealing with laws about the distribution of sugar water. We're dealing with copyright laws in the context of a work which was designed to include material licensed under the GPL as well as under the QPL. And yet you continue to hold that it's just a coincidence that a running copy of kghostscript is going to just happen to include the QPL licensed material. > That's why when you distribute a dynamically-linked kghostview > you don't have to distribute Qt source code. Now of course this > executable work no more contains Qt if it is distributed on the same > CD. (However, in case it is on the same CD, the "special exception" > would not apply.) I presume that by "contain" you're not refering to any sort of physical topology. After all, it's just bits represented on a piece of plastic. There's no "inside" or "outside" in the physical sense. You've got a few bits over here which happen to represent (to an informed person) this concept of "a filename". Near those bits are a few more bits that happen to indicate some other region of bits which happen to represent "the contents of a file". And, on that CD, there are literally thousands of these collections of bits which we think of as "files". But, the GPL doesn't make any sort of claim that only a single file is considered to be the program. Whenever it refers to files, it's very clearly refering to multiple files. But somehow you've gotten this idea in your head that the program kghostscript is exactly the same thing as the file which happens to have the name kghostscript. But you've never offered any evidence for that belief. Please present some evidence that that one file represents the whole program. -- Raul