Jeff, please *don't* put words in your mouth... did I say you were naive? or that you were in an ivory tower? My parents were academics... both professors... there's a big difference between having what I see sometimes as the luxury of being in academia (I really like the spirit of cooperation and collaboration that is more frequent in academia), and the dog-eat-dog competitiveness outside in the "real" world.
Please go look at this link, it will probably explain things better than I can... I don't seem to be all that good at getting my points across! http://www.tokutek.com/tokumx-for-mongodb/ TokuMX is an open source, high-performance distribution of MongoDB that > dramatically improves performance and operational efficiency compared to > basic MongoDB. TokuMX is a drop-in replacement for MongoDB and reduces > costs associated with development, scaling, and optimization: > > 50x performance improvements > 90% reduction in database size > Support for ACID transactions and MVCC > The Tokutek distribution has exactly the same language interface as > MongoDB so you can use TokuMX without changing your application, but it > replaces 1970s B-tree indexing with modern Fractal Tree® indexing > technology. The result is a distribution of MongoDB that greatly > accelerates performance <http://www.tokutek.com/mongodb-performance>, > dramatically cuts disk and flash-drive storage > <http://www.tokutek.com/mongodb-compression>, and supports full ACID > transactions <http://www.tokutek.com/mongodb-transactions>. An old friend, classmate, and 6.111 lab partner was one of the founders... I think he even works in the same building as you! (Brad Kuszmaul... great guy) On Friday, May 1, 2015 at 12:06:57 PM UTC-4, Jeff Bezanson wrote: > > Look, to me, "drop-in replacement" means something very specific that > even a naive academic can understand. It means I can switch "using > MongoDB" to "using TokuMX" at the top of my code, change *nothing* > else, and have everything work. I might speculate, from my ivory > tower, that this property is relevant when trying to get customers to > switch to your product. > > On Fri, May 1, 2015 at 10:35 AM, Scott Jones <[email protected] > <javascript:>> wrote: > > > > > > On Friday, May 1, 2015 at 12:10:32 AM UTC-4, Jeff Bezanson wrote: > >> > >> On Thu, Apr 30, 2015 at 5:26 PM, Scott Jones <[email protected]> > >> wrote: > >> > Maybe because it seems that a lot of the major packages have been put > >> > into > >> > Base, so it isn't a problem, as MA Laforge pointed out, leading to > Base > >> > being incredibly large, > >> > >> That's absurd. There are 500 packages. We added Dates and...what else? > >> We would like Base to be a bit smaller > >> (https://github.com/JuliaLang/julia/issues/5155), but "incredibly > >> large" is a bit of an overstatement. It's *nothing* compared to > >> matlab's default namespace for example. > > > > > > 1) Anything with a GPL license... that is really nasty to anybody would > like > > to use Julia on a commercial project... > > (I have nothing against GPL, I do like OSS, but I much prefer the > way > > the MIT license works, and don't have the luxury of being able to use > GPL > > software in what I do for a living... [use in the sense of using a > library, > > if it's not under the LGPL]... I *use* a lot of GPL'ed software, Emacs, > gcc, > > ...) > > > > 2) multimedia.jl, linalg.jl, statistics.jl, sparse.jl > > [fftw.jl, dsp.jl - I know these are also under my above GPL list, but > if > > a non-GPLed alternative is found, I still think it doesn't need to be in > > "Julia-lite"] > > quadgk.jl, profile.jl, Dates.jl > > > > pkg.jl I'm not sure about... you'd need a way of loading it, to load > other > > packages, but... doesn't it use GPLed software, which could get people > using > > it into legal hot water? > > > > I'm not sure about: reducedlm.jl, combinatorics.jl, don't know what they > do, > > or how basic their functionality is... > > > > Also Markdown, I think there is a lot there that isn't needed just for > > getting ? help documentation (or with @doc) at the terminal... > > Anything not needed for @doc, I think should be optional, in a package. > > > > > >> > with stuff that means Julia's MIT license doesn't mean all that much, > >> > because it includes GPL code by default... > >> > >> So the license of the entire compiler, runtime, and 90%+ of the > >> standard library doesn't "mean much"? Ouch. > >> In any case Viral started adding a flag to exclude GPL libs last week. > >> The changes for that are tiny. > > > > > > Yes, and I'm very grateful to Viral, because otherwise we'd probably > have > > had to totally stop planning on using Julia... > > However, I feel that the developers should be very careful to not let > GPL > > get into the base distribution... > > (I think the default for 0.4 release should be without the GPL > encumbered > > parts) > > > >> > >> I'm still confused about MongoDB vs. TokuMX. In your last post about > >> them you mentioned using them as drop-in replacements for each other. > >> But before that you said they are competitors, and won't necessarily > >> implement the same interface. If they have incompatible interfaces, > >> how can they be drop-in replacements? I don't get it. > > > > > > Do you remember the lawsuits about Java vs. Microsoft's version of Java? > > Or go look at the cringing README.md for the matlab compatibility > package > > for Julia... > > Think about how AMD and Intel battled over extending the x86 instruction > set > > from 32-bits to 64... > > Intel's approach was to introduce the Itanium chip... (real winner > there! > > ;-) We called it the Titanium chip... going down like the Titanic!) > > AMD went and extended the x86 instruction set... then Intel went back > and > > introduced a new instruction set that was mostly compatible with the AMD > > 64-bit instructions... > > That is life outside of academia! > > > > TokuMX recreated the MongoDB's API... but that doesn't mean that > MongoDB's > > developers are going to stop adding new things (sometimes precisely in > an > > attempt to lock people into using MongoDB, make it harder to switch to > some > > other platform), or that TokuMX hasn't added it's own new things. > > I had a part in playing this sort of game for years... with multiple > vendors > > of an ANSI standard language, each adding their own extensions, > sometimes > > having those extensions copied by other competitors... > > > > Scott >
