From: Steve K. <st...@st...> - 2004-03-06 21:42:17
|
On Mar 6, 2004, at 1:22 PM, James H. Cloos Jr. wrote: >>>>>> "Steve" =3D=3D Steve Kann <st...@st...> writes: > > Steve> 2) I _think_ iLBC's license is also some kind of modified BSD > Steve> license. I'm pretty sure that it isn't GPL, and since it was > Steve> presumably compatible with the GPL for asterisk, it's probably > Steve> LGPL compatible for us to use as well. > > Actually, unless they gave Digium different terms than the ones on > their web site, ilbc is unfortunately not gpl or lgpl compatable. > > There is a clause that says you cannot use the code except to create > or decode iLBC bit streams, and said bit streams must be exactly the > same as what the reference code does. That clause makes compatability > with the gpl, lgpl and some other floss licences=B9 impossible. > > Were they do follow Xiph's example and license the code and the > trademark under separate terms, freeing the code but requiring > the current license's bit-stream compatability to use the iLBC > trademark, things would be much better indeed. We'd also be > able to improve the reference code's speed by adding support > for the various simd intruction sets, et al. Hmm, maybe Mark can chime in on asterisk licensing. It sounds like it=20= would be GPL-incompatible. However, even totally proprietary stuff can be LGPL-compatible; LGPL=20 allows you to link it to code under more restrictive licenses. So, you could, for example, use LGPL iaxclient with iLBC and a=20 proprietary, BSD, or LGPL-licensed UI (assuming you made the source=20 available to the LGPL portion, and linkable binary object code=20 available for the non-LGPL portions), but you could not use it with a=20 GPL-licensed UI. |