On 6/6/08, Alessandro Zummo <azummo-lists at towertech.it> wrote: > On Fri, 6 Jun 2008 11:10:52 -0400 > > "m. allan noah" <kitno455 at gmail.com> wrote: > > > > yes- this seems reasonable, however, this 'program' cannot be derived > > from existing GPL'd software that does not already have this added > > permission, because that would change the original program's license > > without permission of the authors. > > > correct. > > > > so, is our answer to Mengqiang that there are only four choices? > > > > 1. you can write an entirely free backend, and use code from SANE. > > > yay! :) > > > > > 2. you can write a partly free backend, that runs the closed parts as > > a separate process, and use code from SANE in the free part, provided > > that the interface to the closed parts is simple and well documented. > > > this is the hp way if i've got it corretly. I find it a bit ugly. > > > > > 3. you can write a partly free backend, that dynamically links to the > > closed parts, provided that you place a license exception in the free > > part allowing said linking. you cannot use any code from SANE, other > > than sane.h and the sane specification, in either part. > > > this means that the sane I/O facilities cannot be used. however > it may be the cleanest thing. > > that's similar to the epkowa way, which uses sane io facilities > iirc?
well, if epkowa dynamically links and uses sanei, then it is not using #3- it might be violating the license? Olaf- can you describe the mechanism? allan -- "The truth is an offense, but not a sin"
