Ps that said, spark community is very accomodating at fixing things i needed for Mahout integration. They seem to have a great momentum, a particulary easy and quick contribution cycle, so things are quick and easy to get done with them. On Oct 7, 2013 12:08 AM, "Dmitriy Lyubimov" <[email protected]> wrote:
> > On Oct 6, 2013 7:20 PM, "Ted Dunning" <[email protected]> wrote: > > > > Yes. Indeed. > > > > Dmitriy has been working with them. I was referencing them when I > > mentioned MLI (their super high level system based on mllib) > Unfotunately i did not get much traction there. They are taking willing > contributers but as far as Mahout is concerned i did not get anywhere. I > think Mahout has an edge in some areas but it clearly has weaknesses in > terms of iterativity, interactivity, environment for some custom things and > pipeline driving. My biggest hope id that scala, a proper dsl and spark > with graphX will remove those from consideration, leaving just the edge. > > > > > > > On Sun, Oct 6, 2013 at 4:21 AM, Sebastian Schelter <[email protected]> > wrote: > > > > > I think we should also try to talk to the mllib guys from the Spark > > > project and see if there's some potential for collaboration. > > > > > > Best, > > > Sebastian > > > > > > > > > On 06.10.2013 13:18, Isabel Drost-Fromm wrote: > > > > On Saturday, October 05, 2013 01:11:40 PM Ted Dunning wrote: > > > >> I was asked to answer an anonymous question about the future of > Mahout > > > on > > > >> Quora and thought I should share the answer here as well. > > > > > > > > > > > > Great summary indeed. Thanks for sharing Ted. Agree with all of your > > > points. > > > > The only thing I'd add would be a documentation clean-up, but that's > > > something > > > > that could get some attention with moving to Apache CMS. > > > > > > > > > > > > Isabel > > > > > > > > > > >
