On 07/12/11 08:33, Leo Simons wrote:
On Wed, Dec 7, 2011 at 1:24 AM, Benson Margulies<[email protected]> wrote:
On Tue, Dec 6, 2011 at 7:03 PM, Leo Simons<[email protected]> wrote:
More legal stuff...on reproducing licenses and notices....
Since The Official Release is the source, and the binary releases are
'just for convenience', do we really need to apply all the same
fine-tooth-comb treatment?
Of course! A lot of open source licensing is about "things you have to
do when you (re)distribute".
In a binary release with libraries inside, apache is (re)distributing
those libraries, so those license terms kick in, and we simply have to
follow them. And a lot of the obligation that comes with those
licenses is about reproducing licenses and notices.
The only alternative is not to redistribute those libraries, which
puts the burden of figuring out the legal mess on the user (and so
makes the binary "less convenient" for them!).
Finally, do remember that _normally_ you would update the binary
release LICENSE and NOTICE when you updated its build scripts to
include another dependency, so this painful legal cleanup/scrubbing
should really be a once-only event.
I'm more than happy to try to get the details right - I'm not aiming to
merely meet the minimum necessary requirements but to do the "right
thing" now. We're learning.
Jena has historically shipped with dependencies, originally (10 years
ago) because that was the only practical way to get the stuff to the
users. The simple act of putting the right versions of the software
into the users hands was been quite important. Getting the classpath set
correctly is another matter ...
Nowadays, because public maven, or even maven, isn't for everyone, it's
still useful to deliver the bundle. I guess all of us use maven (or
etc) dependency management but it's a real barrier for people starting
out with Java, semantic web all at the same time, and not just students
doing projects. Ditto people who use the command line tools.
So I appreciate the fine tooth combing. At the moment I'm on the
receiving end of formal public comments on specs as SPARQL goes through
the later stages of W3C process, and the comments there are not
necessarily as supportive as here.
Andy
