From: Richard S. <ri...@ex...> - 2009-04-19 18:55:12
|
[Have just got back from a few days away and am catching up] Don Morrison wrote: > Others may think of other points, but I think the major consideration > here is > > If we stick with the GPL, folks that sell proving software and so > on will not be able to incorporate our stuff unless they make > their own sources freely available, which I presume they'd be > loath to do. The counter argument is: why I / we / Graham put a lot of effort into developing the application and give it away for free without getting anything back in return? By GPLing it, if someone creates a derivative program, we'll be able to use their work in future programs that we might write. I would certainly prefer to see anything clever that I contribute significantly to (whether client or server side) under a strong copyleft licence such as the GPL. If it ends up that the server and/or default client aren't all that clever, then I'm less concerned. (For example, the methods database server code isn't all that complicated, and, although we released it under the GPL, I wouldn't be concerned if we changed this to something more perrmissive. And the C++ client code is already under the more permissive LGPL.) So I'd be quite content to leave it as GPL at the moment and, if part of it turns out to be fairly trivial, we can relicense it later. And as Graham says, none of this stops us from documenting the wire protocol and allowing anyone to interface to this without restriction. RAS |