In praise of Open - Proprietary hybrid software releases
In my post Why Open Source Software? here is what I wrote in the end paragraph
Let's see for example The R Project for Statistical Computing . Just look at one field High-Performance and Parallel Computing with R there is a mind-blowing large lists of packages available. The documentation is cryptic, as the authors assume everyone have the same abilities to understands as them. On the other hand, the R open source ecosystem has created a wealth of statistical applications impossible to be covered by proprietary software. Read Will 2015 be the Beginning of the End for SAS and SPSS? This is already happening as we learn from this tweet
This is probably a classic example of hybrid Open and Proprietary software release
Google has released three weeks ago two R packages
But how to get to this stage? The primary function of the proprietary partner in the Open-Proprietary model is to discover a sustainable (and scalable) business model. This could never be done in open source, where the development looks like this:
Steve Blank writes:
Startups from open source are not the only option
Open Source Software is beneficial... Most of us, we do want to do goodness. There is no way to accomplish our mission and keeping the freedom, if we starve to death, if we don't have children (like Richard Stallman), if we don't get funding for our projects and if no one uses what we invented and created ex-nihilo.Danese Cooper, ex Sun, aka the Open Source Diva is a follower of Meher Baba, the Irani Indian spiritual teacher. She is the most influential, sensitive and inspired voice and advocate of Open Source. This is what she says on background page of her LinkedIn profile:
I believe that the Open Source Methodology of increased transparency in software development (and possibly also in hardware development) can only have a positive effect on the technology we deliver to customers. The marriage of Open Source to traditional software development has created some interesting hybrid (part Open part Proprietary) models which I believe hold up favorably against entirely traditional methods in both time to market and satisfaction of market demand.Danese means here something different than simply offering a model of free code and charging for support. Or as we did in Sun Micro with Sun Grid Engine, we offered two distributions, one with a priced license and one for free (using the most liberal SISL open source license ). The latter proved a failure, but this is another conversation
Let's see for example The R Project for Statistical Computing . Just look at one field High-Performance and Parallel Computing with R there is a mind-blowing large lists of packages available. The documentation is cryptic, as the authors assume everyone have the same abilities to understands as them. On the other hand, the R open source ecosystem has created a wealth of statistical applications impossible to be covered by proprietary software. Read Will 2015 be the Beginning of the End for SAS and SPSS? This is already happening as we learn from this tweet
Revolution Analytics VP Marketing, David Smith reports that their "development team spent more than six months making R work on 64-bit Windows (and optimizing it for speed), which we released as REvolution R Enterprise bundled with ParallelR" They call it an R Enterprise grade and it satisfy the desires of their investors - which include Microsoft and Intel - to use the Windows 64 bit operating system. According to a 2011 press release, the prices start at $1,000 per single user.Top / All / People you follow
IBM division switches from SAS to R to save money, partly to avoid charges from Global Services ... Ironic, eh? #dataweek
Danese Cooper This is probably a classic example of hybrid Open and Proprietary software release
Google has released three weeks ago two R packages
Both Revolution Analytics and Google do not develop from scratch. They adapt from a wealth of open source packages software releases that meet the needs of their own customers, financial backers or speed the internal development of new productsRProtoBuf is an R package for Google’s Protocol Buffer library that allows one to define simple data structures with intuitive getter and setter methods.HistogramTools is a new R package that uses RProtoBuf to read in a compact protocol buffer representation of binned data
Why a mix of open and proprietary software development ?
The Agile Manifesto is quoted by Tristan Kromer . He quotes
"Our highest priority is to satisfy the customer through early and continuous delivery of valuable software."
But what is a valuable software? Tristan answer is
"Value is determined by the customer in the act of paying for a product."Open source releases have no monetary value. Proprietary established companies can re-use open source in a customer environment where there are people, aka customers, willing to place their hands inside their pockets, take our their valets and pay.
But how to get to this stage? The primary function of the proprietary partner in the Open-Proprietary model is to discover a sustainable (and scalable) business model. This could never be done in open source, where the development looks like this:
 |
| From Tristan Kromer: Open Source like Agile Develoment |
Scientists as entrepreneurs
Yes. Traditionally entrepreneurship and science are compared to oil and water.
Steve Blank and Bob Dorf change this perception. For 3 years in row they teach the National Science Foundation Innovation Corps (I-Corps) - professors and their graduate students - how to bridge the commercialization gap between their scientific research and wide-scale commercial adoption and use. Quote:
From my experience in working with extremely talented scientists, a startup is a harder sell solution for a professor or scientist. My comment to Steve Blank blog on LinkedIn is:The program pays scientists $50,000 to attend the program and takes no equity.Currently there are 11 U.S. universities teaching the Lean LaunchPad curriculum organized as I-Corps “nodes” across the U.S. The nodes are now offering their own regional versions of the Lean LaunchPad class under I-Corps.The NSF I-Corps uses everything we know about building Lean Startups and Evidence-based Entrepreneurship to connect innovation to entrepreneurship.
I preach for over a year some initiatives to take research ideas to commercialization, Going "out of the door" and speaking with customers is adverse to most scientists psychology. It belittles lofty aspirations and researchers lack the risk taking capability. It is doable, but the devil is in the detailsIn my humble opinion Hybrid Open - Proprietary commercialization will have much greater chances to success starting from scratch a startup. Researchers and professors, can keep their academic position and research staff, while the risks are mitigated by commercial companies.
Steve Blank writes:
Every time I go to Washington and spend time at the National Science Foundation or National Institute of Health I’m reminded why the U.S. leads the world in support of basic and applied science. It’s not just the money we pour into these programs (~$125 billion/year), but the people who have dedicated themselves to make the world a better place by advancing science and technology for the common good.Here is the progress report from Universities of Stanford, Berkeley and UC San Francisco. Jerry Engel was my professor at UC Berkeley Entrepreneurship program
Startups from open source are not the only option
Hybrid Open Source - Proprietary projects
These projects may create jobs for the same open source contributors who are now laboring for free
As a non-cliche Product Manager here are some ideas For each idea listed here, there about 100 more ideas to discover.
- Build access to any cloud or grid using a laptop (Amazon Web Services, Joyent Manta, etc)
- Build access to any TOP 500 supercomputer using a laptop
- Commercialize an R package based in Bosco GridR to distribute the huge R jobs to all nodes for maximum throughput in a cloud
- Submit performance jobs via email. Yes, email
- More suggestions supplied on request
Does any corporation have a need for these type of applications? Or, alternatively, describe your problem. Interested?
Comments