Adam Warner wrote: >PPS: Does anyone know of GFDL-licensed documentation that contains >significant verbatim quotations of GPLed code?
Yes. The libstdc++ manual is largely generated from Doxygen comments in GPL'ed files. (But there are some skeleton files which are straight GFDL, so it can't be purely GPLed.) I have raised the licensing issue with the GCC mailing list [EMAIL PROTECTED] >As an idea this GPL code has been taken from Emacs: >http://www.gnu.org/manual/elisp-manual-21-2.8/html_node/elisp_350.html >About 140 lines of code have been copied from a 207 line file. We know >the licences are incompatible. Is this OK without the Free Software >Foundation explicitly dual licensing the code? (I'd say yes this is OK >because all the code is assigned to them so they can do what they like, >while still maintaining that we can't do the same). Right. The FSF can distribute this code; it's dual-licensed simply because they, the copyright holder distributed it under both licenses. :-) If anyone else changed the GPLed or GFDLed copy and didn't explicitly dual-license, nobody could port those changes to the other copy (except the copyright holder). Nobody but the FSF can legally distribute a revised or amended version of the libstdc++ manual regenerated from source, because the regeneration process involves taking text from files licensed under the GPL and putting it in a document licensed under the GFDL. (If the skeleton files were replaced with GPLed skeleton files, then a GPLed manual could be created.) >This at least demonstrates that the situation is unworkable without >being in the privileged position of owning all copyrights. Yep. -- Nathanael Nerode <neroden at gcc.gnu.org> http://home.twcny.rr.com/nerode/neroden/fdl.html