On Fri, Nov 25, 2005 at 02:56:02PM -0500, Glenn Maynard wrote: > On Fri, Nov 25, 2005 at 07:23:24PM +0000, Måns Rullgård wrote: > > >>> > Do you think that this licence does not require a developer > > >>> > of a modified package (other than PHP) to lie by saying > > >>> > "This product includes PHP software"? > > >>> > > >>> Perhaps the PHP folks subscribe to the view that PHP scripts are > > >>> derivative works of PHP. > > >> > > >> Ye ghods, I'd hope not. That would be similar to believing that this > > >> message is a derivative of the English Grammar manual I read in school. > > > > > > Or that all non-trivial Emacs Lisp code must be licensed under the > > > GPL. This position is not *that* unusual... > > > > Not being unusual doesn't make it sensible or correct. > > Just to take a guess at where this strange claim might have originated: > > The FSF (from what I understand) claims that binaries linked against GPL > libraries are derivative works of the library, because the resulting > binary has pieces of the GPL software in it. This isn't obviously true > with C libraries, which has led to a lot of debate around the topic, but > the claim isn't entirely unreasonable.
Assuming that "linked" in your paragraph above means "dynamically linked" (as your second sentence suggests), can you provide a cite from the FSF which makes this claim, with rationale? I looked around, as research for my blog post "Linking does not create a derivative work" (http://www.hezmatt.org/~mpalmer/blog/general/linking_does_not_create_a_derivative.html) and couldn't find anything that really actually made the claim in those terms. - Matt
signature.asc
Description: Digital signature
