Richard, No, this change should not affect you at all. --------- Matt Taylor OS Community Flag-Bearer Numenta
On Sun, Jul 19, 2015 at 7:47 PM, email email <[email protected]> wrote: > Hi, Matthew, > > I would like to know if the license changes research students or fellow like > me will not download and study the source code? or develop our own research > programs on the base of NuPIC? > > Thanks for your response > > Richard > > On Sun, Jul 19, 2015 at 1:30 AM, Fergal Byrne <[email protected]> > wrote: >> >> My view as someone working to make a living using and extending HTM >> technology is very simple. If you stay in (A)GPL-land, you're completely >> free. If you build something for private use or on behalf of a single >> client, you're good (you may transfer the responsibility to your client if >> they're making money). Otherwise contact Numenta. >> >> There will be billion dollar businesses based on HTM in the next few >> years. Those businesses should repay Jeff and Donna for the money and effort >> they invested to start this. >> >> There is huge debate in Deep Learning right now about Google's decision to >> try and patent several key ideas, many of which go back decades. So it's not >> just us. >> >> Regards >> >> Fergal Byrne >> >> -- >> >> Fergal Byrne, Brenter IT >> >> Author, Real Machine Intelligence with Clortex and NuPIC >> https://leanpub.com/realsmartmachines >> >> Speaking on Clortex and HTM/CLA at euroClojure Krakow, June 2014: >> http://euroclojure.com/2014/ >> and at LambdaJam Chicago, July 2014: http://www.lambdajam.com >> >> http://inbits.com - Better Living through Thoughtful Technology >> http://ie.linkedin.com/in/fergbyrne/ - https://github.com/fergalbyrne >> >> e:[email protected] t:+353 83 4214179 >> Join the quest for Machine Intelligence at http://numenta.org >> Formerly of Adnet [email protected] http://www.adnet.ie >> >> >> On Sat, Jul 18, 2015 at 6:21 PM, Austin Marshall <[email protected]> >> wrote: >>> >>> I considered the MongdoDB example. One interesting aspect of MongoDB is >>> that while the server is AGPL, the drivers are Apache. This helps to make >>> it straight-forward from a use/licensing standpoint. If I create mongodb++ >>> as you describe, I'm bound to AGPL, but if I run it verbatim, and connect to >>> it with the official drivers, then I'm not bound by AGPL and am free to >>> build a proprietary product. I don't think such an arrangement is being >>> suggested here. >>> >>> The pro-forma idea is interesting, and certainly helps to mitigate the >>> barriers for entry for commercial applications. >>> >>> On Sat, Jul 18, 2015 at 10:05 AM, Fergal Byrne >>> <[email protected]> wrote: >>>> >>>> Hi Austin, >>>> >>>> The best example of this is MongoDB, which operates under the AGPL for >>>> this reason. I cannot take MongoDB on its own, modify it in some way (let's >>>> say by incorporating a new caching algorithm) and then market it as my own >>>> MongoDB++. I also couldn't provide MongoDB++ as a service without providing >>>> full source to my mods. >>>> >>>> I also raised your other point with Donna, about using NuPIC as a back >>>> end for some software service. As far as I can tell, you could simply build >>>> an API layer on NuPIC and release only that under (A)GPL, while keeping >>>> your >>>> client app completely closed. Since your code is only a "user" of NuPIC, >>>> this would fall outside the AGPL per se. You would however then be in the >>>> territory of building technology using Numenta-patented IP, so it'd be wise >>>> to think about talking to Donna at that point. >>>> >>>> On your comments re bootstrapping to an MVP using NuPIC, would it make >>>> sense for Numenta to develop a "pro-forma" precommercial license, with a >>>> timeout such as 12 months? This way, you can notify Numenta that you are >>>> likely to want to negotiate a commercial license, but they give you a year >>>> to get to that point at no cost. In the event you get no traction, just >>>> Open >>>> Source your code and you're good. >>>> >>>> Regards, >>>> >>>> Fergal Byrne >>>> >>>> -- >>>> >>>> Fergal Byrne, Brenter IT >>>> >>>> Author, Real Machine Intelligence with Clortex and NuPIC >>>> https://leanpub.com/realsmartmachines >>>> >>>> Speaking on Clortex and HTM/CLA at euroClojure Krakow, June 2014: >>>> http://euroclojure.com/2014/ >>>> and at LambdaJam Chicago, July 2014: http://www.lambdajam.com >>>> >>>> http://inbits.com - Better Living through Thoughtful Technology >>>> http://ie.linkedin.com/in/fergbyrne/ - https://github.com/fergalbyrne >>>> >>>> e:[email protected] t:+353 83 4214179 >>>> Join the quest for Machine Intelligence at http://numenta.org >>>> Formerly of Adnet [email protected] http://www.adnet.ie >>>> >>>> >>>> On Sat, Jul 18, 2015 at 5:33 PM, Austin Marshall <[email protected]> >>>> wrote: >>>>> >>>>> I see a couple of issues with the AGPL. The wording of AGPL is >>>>> absolute regarding "all users interacting with it remotely through a >>>>> computer network". This may become a barrier for anyone potentially >>>>> interested in a commercial license later. For example, in Lean Startup >>>>> parlance, I may want to experiment with a minimum viable product to gauge >>>>> interest before investing too much effort into developing a commercial >>>>> product. If I'm not already planning on using AGPL (very few do), I'm >>>>> forced to consider the implications before I start my experimentation. >>>>> Personally, I'd rather not have to worry about it -- I'd want to get users >>>>> on my mvp as early as possible and not have to delay the process with >>>>> commercial license negotiations, especially since I'd be at a >>>>> disadvantage, >>>>> not having any experience with the technology and not having much of an >>>>> opportunity to make an informed estimate of the commercial viability. >>>>> Google has even taken a stand and outright banned AGPL-licensed software >>>>> for >>>>> internal use >>>>> (http://www.theregister.co.uk/2011/03/31/google_on_open_source_licenses/). >>>>> I find that position to be reasonable, and I'm sure they are not alone. >>>>> >>>>> I'm also skeptical about the enforceability the AGPL with respect to >>>>> closing the loophole. Let's say I want to get my product out and either >>>>> don't want to seek out a commercial license or don't want to do it now. I >>>>> might argue that, in many ways, you can incorporate nupic into your tech >>>>> stack and not be required to share your source. For example, let's say I >>>>> have a product that makes recommendations, and behind the scenes I use >>>>> nupic >>>>> in some small part of an ensemble. If this process is done in an >>>>> offline/batch mode on behalf of the user and only the results conveyed to >>>>> the user, then I might argue that my user has no interaction with nupic, >>>>> and >>>>> therefore my service is not subject to the virality of the AGPL. >>>>> >>>>> In other words, I'm either likely to avoid it like the plague, or try >>>>> to get crafty. I'm not convinced that the AGPL helps in the long term >>>>> adoption of nupic and related software, from either free or commercial >>>>> licensing standpoints. >>>>> >>>>> On Fri, Jul 17, 2015 at 10:00 AM, Matthew Taylor <[email protected]> >>>>> wrote: >>>>>> >>>>>> Good questions... >>>>>> >>>>>> On Fri, Jul 17, 2015 at 9:46 AM, Dean Horak <[email protected]> >>>>>> wrote: >>>>>> > How does a project transition from GPLv3 to a different license, >>>>>> > when all the existing code has already been released as GPLv3. I >>>>>> > assume that >>>>>> > the GPLv3 license will remain in effect for all existing code, and >>>>>> > only new >>>>>> > code specifically contributed by Numenta specifically identified as >>>>>> > AGPL >>>>>> > will be affected by this. >>>>>> >>>>>> Correct. The NuPIC (and related) code that currently exists on Github, >>>>>> and all the history of that code, is GPLv3. There is nothing we can do >>>>>> about that, it will always be GPLv3. When we change the license to >>>>>> AGPLv3, from that point forward, the repository and all future >>>>>> developments in the repo will be AGPLv3. So there will be a line drawn >>>>>> in time at the commit SHA when we make the license change. >>>>>> >>>>>> > But what about community contributed code? Surely Numenta cannot >>>>>> > force the community to adopt AGPLv3 should they choose not to since >>>>>> > Numenta >>>>>> > is technically only a contributor (albeit the prime contributor) as >>>>>> > well and >>>>>> > not the "owner" of the codebase. >>>>>> >>>>>> Actually, Numenta is the sole copyright owner of the NuPIC codebase, >>>>>> and the copyright owner has control over the license of the code. This >>>>>> means that Numenta, as the copyright owner, has the legal right to >>>>>> change the license without input from any contributors, because all >>>>>> contributors signed our Contributor License Agreement [1] that signs >>>>>> over all their copyright of their contributions to Numenta. >>>>>> >>>>>> > Do contributors have the option of choosing AGPLv3 or GPLv3? I >>>>>> > suppose a vote from the community to adopt AGPLv3 for all future code >>>>>> > could >>>>>> > be enforced by the committers - only allowing AGPLv3 code into the >>>>>> > codebase, >>>>>> > but this seemingly could lead to a fork of the code, which is probably >>>>>> > not a >>>>>> > desirable outcome at this point. >>>>>> >>>>>> No, contributors will not get a choice in the matter. If this codebase >>>>>> were copyright many authors, a vote would be necessary to change the >>>>>> license. But because Numenta is the sole copyright owner, a vote is >>>>>> unnecessary. We do, however, care what our contributors think about >>>>>> licenses, and we certainly to not make such changes wantonly. >>>>>> >>>>>> > Again, I do not expect that this will have any real impact on me, >>>>>> > but in the spirit of clarity and transparency, I think response to >>>>>> > these >>>>>> > types of questions should be considered. >>>>>> >>>>>> I am happy to answer any more questions. >>>>>> >>>>>> [1] http://numenta.org/licenses/cl >>>>>> --------- >>>>>> Matt Taylor >>>>>> OS Community Flag-Bearer >>>>>> Numenta >>>>>> >>>>> >>>> >>> >> >
