mahout-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ted Dunning <ted.dunn...@gmail.com>
Subject Re: Commercializing Mahout: the Myrrix recommender platform
Date Wed, 04 Apr 2012 21:56:15 GMT
The answers are wildly different.

One of the problems is that if the code line has diverged significantly, it
can take a lot of work to build up a patch.

What we do at MapR is try to draw a few lines.  In the outer circle, we try
to run as close to the released code as possible.  This makes contributing
fixes relatively easy.  In the middle zone, we try to maintain coherence in
spite of large changes so that we can contribute patches where they can be
applied to the OSS versions.  In the core, there is no corresponding OSS
code in any case so it doesn't matter.

Feature requests fall into these same categories depending on what they
apply to.  If secret sauce is essential, then the resulting code probably
can't reasonably be released.  Or more correctly, the code probably can't
be accepted by the OSS project since most users wouldn't be able to run it.
  The exception would be features that are protected by config switches
that default to the safe position.

On Wed, Apr 4, 2012 at 2:43 PM, Dan Beaulieu <danjacob.beaulieu@gmail.com>wrote:

> I've always wondered how businesses who provide their own solutions using
> OSS decide when to contribute code back to the project.
> For instance, suppose a bug is logged, and the business already has a fix
> in their fork. Do they provide a patch?
> What about a feature request? What if the deliverables for the feature
> includes secret sauce from the business.
>
> Maybe this isn't really a common occurrence, or not how the OSS business
> model works, but figured it's a good as time as any to ask I figure.
>
> On Wed, Apr 4, 2012 at 5:35 PM, Ted Dunning <ted.dunning@gmail.com> wrote:
>
> > With this announcement, this group has a fork in the road facing us.
> >
> > We can choose the Hadoop path of forcibly excluding anybody with a
> slightly
> > wrong commercial taint from discussions (I call this the "more GNU than
> > GNU" philosophy).
> >
> > Or we can choose a real community based approach that includes vendors
> > regardless of how they use the code that we freely give away via the
> Apache
> > Mahout project (I call this "the Apache way").
> >
> > As you may guess from the way that I phrase these options, I would prefer
> > the second approach.
> >
> > As such, I like it if we could resolve as a group that we very much
> welcome
> > what Sean is doing as an augmentation rather than diminution of the major
> > role that he has played in Mahout so far.  More than that, I would like
> to
> > go on record saying that I, at least, am happy to have all kinds of
> > participation in Mahout.
> >
> > Is this the consensus here?  I think it is important to bring this
> subject
> > up early and get a definitive consensus rather than let it drift.
> >
> >
> >
> > On Wed, Apr 4, 2012 at 12:33 PM, Sean Owen <srowen@gmail.com> wrote:
> >
> > > Dear all -- I've long promised (threatened?) to begin efforts to
> > > commercialize Apache Mahout. Given my line of work in VC, I see
> > > evidence for positive symbiosis between open source and commercial
> > > enterprise. We have evidence from the growth in user base and mailing
> > > list, as well as the Mahout in Action book, that this is becoming
> > > another successful Apache project. It's time to start bringing more of
> > > Mahout to the commercial world, and in turn bring those benefits back
> > > into the open-source project.
> > >
> > > To that endm I've been working on a heavily product-ized, somewhat
> > > evolved, version of the recommender engine code available in Mahout.
> > > After about six months of work, it's ready to announce. Please meet a
> > > new startup company and software product, Myrrix (http://myrrix.com).
> > >
> > > Well, at least, it's ready for pre-launch of part of the product. The
> > > announcement (http://myrrix.com/myrrix-pre-launches/ ) tells part of
> > > the story, and the design doc (http://myrrix.com/design/ ) tells more,
> > > but I'll summarize key points here:
> > >
> > > - Aims to add packaging, documentation and support, to product-ize
> > > - Reuses Mahout APIs, some code, and certainly presented as a
> > > "Mahout-based" platform
> > > - Two-tier architecture: serving (or "speed") layer acting in
> > > real-time, coordinating with distributed Hadoop-based computation
> > > layer. Best of both worlds.
> > > - Leans heavily on a variant of the alternating least squares algorithm
> > > - Serving layer is free/open source and is a complete solution for
> > > small-to-medium recommender problems
> > > - Full distributed architecture, including hosted offering, is the
> > > commercialized (read: for-pay) part
> > >
> > > I would welcome community support in telling me what you like and
> > > don't like about this. All the better if anyone is in a position to
> > > actually run and test the Serving Layer: http://myrrix.com/download/
> > >
> > > I intend this to bring more profile, attention, and work to the Mahout
> > > project, rather than subtract from it. I hope to see more
> > > Mahout-related commercializations, beyond the inclusions in
> > > distributions we're already seeing, in 2012, as it's key to the
> > > long-term project health. It's most certainly going to be the year of
> > > the application layer (analytics, machine learning) for Big Data.
> > >
> > > Thank you!
> > > Sean
> > >
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message