Steve Holden <[EMAIL PROTECTED]> wrote: > > In the absence of other information I Cc'd [EMAIL PROTECTED] to ask their > opinion about the relationship between GPL'd Python modules and > programs that imported them ... > > If a Python program imports a module licensed under the GPL, in your > > opinion does the Python program become a derivative work of the GPL'd > > > software? > > Generally, yes. > > -- > -Dave "Novalis" Turner > GPL Compliance Engineer > Free Software Foundation > > So, there we have it, not a legal opinion but at least what a GPL > compliance engineer thinks. I must add, I can't remember when I saw a > dafter title.
Definitely not a legal opinion, which are characterised by an explanation of the logic and reasoning behind the opinion. Perhaps you can ask the GPL Compliance Engineer how he responds to the following facts: 0) Section 0 of the GPL states: "Activities other than copying, distribution and modification are not covered by this License; they are outside its scope." 1) In Australia and Europe at least, loading program code from disc into memory in order to execute it is not considered as making an infringing copy under copyright law. Thus it is drawing a very long bow indeed to describe importing a third-party GPLed module at runtime as "copying" in these countries. 2) Importing a third-party GPLed Python module at runtime cannot be considered distribution, nor can it be considered modification, since the source or object code for that module is not being edited or modified. Given these fatcs, under what possible interpretation (in Australia and Europe and probably elsewhere) can the act of importing a GPLed Python module at runtime be considered to fall under the scope of the GPL as set out in Section 0, quoted above? See also this article, in which Lawrence Rosen, the attorney for OSI, admits that it is impossible for anyone to properly interpret the application of various open source licenses under the various national laws: http://www.devx.com/opensource/Article/27171 I suspect that this issue applies equally to the FSF. Tim C -- http://mail.python.org/mailman/listinfo/python-list
