Boost logo

Boost :

From: Michael Stevens (list-boost_at_[hidden])
Date: 2007-05-28 05:25:47


On Sunday 27 May 2007 11:36, Vladimir Prus wrote:
> David Abrahams wrote:
> > on Wed May 23 2007, Janek Kozicki <janek_listy-AT-wp.pl> wrote:
> >> But here's what I've found:
> >
> > <snip>
> >
> >> Whew. That's it. Took me one and a half an hour. Hope this helps.
> >
> > Wow, thanks!
> >
> >> Based on above research I could tell that only boost::ublas doesn't
> >> have BSL, while all the other files are simply a mistake that can be
> >> quickly corrected...
> >
> > In that case I vote for ripping ublas out of Boost unless and until
> > the authors fix it. This is crazy; people have had long enough.

crazy ;-)

> Interesting. Reading http://boost.org/more/lib_guide.htm#License I see
> that BSL is the recommended, but not required license. Above, you propose
> to rip out a part of Boost because it's not BSL. Can you please point me to
>
> - A document that say BSL is an absolute requirement
> - A mailing list announcement that BSL is now an absolute requirement
>
> I've no comment if such change is good or not, but I'm worried about such
> global decision being made silently.

Indeed.

With regard to uBLAS I think there is little chance of a license change.

Joerg was contactable up to last year by phone. Email however landed in the
bit bucket. He was always very non committal as to a change to a BGL license.
I assume a change would require Mathias' agreement. I not sure who Mathias is
or if Joerg was still in contact with him. Basically the orignal authors do
not seem to be interested in a license change.

We would have to accept that if uBLAS's license becomes unacceptable to the
Boost community then uBLAS would have to move. The change to BSL has been
around for a long time after all and the BSL is a good thing.

Maybe there would be a gentler solution then Dave's suggested 'ripping out'. I
think a single license for all of Boost is very helpful to users. So the
change would have to make the license status very clear. Some kind
of 'historical', 'aberrant license' status maybe?

Michael


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk