Archive for April 2010
Interesting post by Matt Aslett at the 451 Group. They surveyed over 1000 open source users globally to find out about the drivers behind open source adoption. Interesting stuff.
I wonder how this would manifest itself on the (theoretical) FOSS/GloMM grid?
A good article by Doug Henschen (I got it right this time, Doug) reviewing part of a presentation from Gartner’s BI Summit last week. Here is Doug’s first statement:
The five-year cost of a typical, 500-seat BI deployment ranges from about $150,000 for an open-source system to just over $1 million for the full suite from SAP BusinessObjects, IBM Cognos or Oracle. In between these two extremes are Microsoft, pure-play BI vendors ($522,000 to $674,000) and software-as-a-service (SaaS) BI vendors ($582,000).
The full article is here: http://intelligent-enterprise.informationweek.com/blog/archives/2010/04/gartners_take_o.html
I have stated a few times that the Free Software Foundation (FSF) and its advocates don’t have a vision of the future that I find viable. I have read statements that the best custodians of FOSS are tiny consulting companies, and that Microsoft and Oracle should be barred from participating in FOSS. I don’t see how the software needs of the world can be met by tiny services companies, or how we can magically make the existing market players disappear. But I can’t complain about their vision without providing any vision of my own. So here it is.
I subscribe to the theory that a vision is a dream + a plan.
Twenty years from now, across the globe, every individual, business, organization, and government entity will have FOSS suitable for all their needs. That is not to say there there is no proprietary software any more – their certainly will be for the next 20 years – just that any and all normal requirements can be met with FOSS.
In this future the notion of intellectual property will still exist, as will software patents (unfortunately). In this future, any software or services company, of any size, whether local or global has the opportunity to participate in the FOSS realm.
We will reach this goal incrementally via an evolution of FOSS software, an evolution of the existing market players, and the creation of new market players.
1 – Establish Metrics
Here is my proposal for assessing the state of the dream. By country, we score each software domain in terms of how well FOSS provides suitable solutions that are:
- OSI approved.
- Compliant with all local regulations (accessibility, domain-specific legal requirements etc).
- Stable, usable, and documented.
- Available on multiple platforms (at least 2).
- Available from 3 separate projects (different code-bases), failing that 3 different distros.
Any software under an OSI license is eligible for inclusion – no matter the size of the project, or the business model of provider.
We also score the software with regards to how well it supports all the needs (including support, training, and professional services) of:
- Micro organizations (1-9 people)
- Small organizations (10-99 people)
- Medium organizations (100-250 people)
- Large organizations (> 250 people)
The domains assessed could be (I’m sure there are many more we can add):
Operating System and Middleware
1. Operating System (OS)
2. Database (RDBMS)
3. HTTP and Application Servers
4. Network Management and Monitoring
5. Enterprise service bus (ESB), message queue (MQ)
7. Instant messaging
9. Customer Relationship Management (CRM)
10. Locally-compliant Enterprise Resource Planning (ERP)
11. Content Management Systems (CMS), knowledge base
12. Call center, case tracking
14. Online meeting and conferencing
15. Voice over IP (VoiP)
16. Collaboration – forums, wiki etc
17. Reporting, analysis and Business Intelligence (BI)
18. Online training
19. Financial, Budgeting and Planning, including public sector
21. Word processing
24. Graphics editors
25. Printing tools
26. Software and web tools (compilers, editors etc)
32. Retail, including Point of Sale
36. Travel and hospitality
37. Engineering, Manufacturing, Construction
Obviously, within each of these vertical domains there are multiple applications. Scoring here will be tricky.
39. The local availability of systems integrators that can implement FOSS stacks and solutions.
Scoring is done per country and per domain and is scored from 0 to 9. A score of 0 indicates there is no FOSS option for that domain and geography. A score of 9 indicates the existence of three different FOSS options that meet the needs of large organizations. We can color code by range, red=2 or less, yellow=3 to 6, green=7 or more
2 – Census
We need to find out how close we are to achieving the dream. Volunteer organizations and sponsoring organizations score each domain for a single country, providing notes about the FOSS packages assessed and any services options assessed. The results of the census are publicly available at all times. Academia and analysts could provide much of this data.
3 – Close Gaps
Based on the results of the census, sponsoring organizations provide resources and guidance to help close the gaps. Sponsoring organizations will have many different motivations:
- Wanting a larger local and global market for their services and support offerings.
- Wanting software that is more accessible.
- Wanting more FOSS options in their country or domain.
- Wanting to sell add-ons and extensions.
Or we just allow the natural progress of FOSS to gradually populate the GloMM in a natural – ‘Game of Life’ / Brownian motion – kind of a way.
4 – Repeat Steps 2 and 3
As time goes by and we repeat steps 2 and 3. The Matrix flushes out gradually, and becomes greener and greener.
5 – Declare Victory
In my opinion FOSS has won when, and only when, the entire sheet (7000-8000 cells) is lit up in green. At this point the value of FOSS will be clear to everyone. Maybe attitudes towards intellectual property will change then. But we can’t expect them to change before we get close to this point.
As this gradual global evolution occurs, the existing market players will have to adapt to new market conditions. What they do, and how well they do it, is up to them – but they are welcome to participate. Just because Oracle is now the ultimate custodian of MySQL, does not mean that MySQL should not be listed as one of the FOSS databases. Microsoft, IBM, SAP, Oracle should be accepted into this evolution – whether they survive it is up to them and the global and local markets, not up to anything else.
Any organization that produces FOSS, or localizations, or documentation, or provides services or support for FOSS is deemed to be a friend of GloMM – no matter what their size, history, or business model.
So that’s my vision. I have a defined goal, a way of measuring progress, mechanisms for getting there, and ways for existing market players to participate. I claim it to be reasonable, rational, and viable.
I have no resources at my disposal to execute on any of this. Its just a vision. If only I had a dream + a plan + resources :-)
Steven Levy, who wrote ‘Hackers: Heroes of the Computer Revolution’ twenty five years ago, has re-interviewed some of the original subjects from the book in a Wired article titled ‘Geek Power: Steven Levy Revisits Tech Titans, Hackers, Idealists‘
Its a long read, but worth it. Very interesting.
There are some illuminating quotes from Richard Stallman – twenty fives years ago he said:
I’m the last survivor of a dead culture. And I don’t really belong in the world anymore. And in some ways I feel I ought to be dead.
In terms of effect on the world, it’s very good that I’ve lived. And so I guess, if I could go back in time and prevent my birth, I wouldn’t do it. But I sure wish I hadn’t had so much pain.
To me, these statements are telling. Why does Stallman feel such pain? I can only think of two possible reasons. Either his ideas and ideals are so far ahead of their time that the world is not ready for them, or they don’t fit well with the world and never will.
His idea that software ‘cannot be owned’ makes no sense to me. The idea that, by sitting at home and writing code on my computer for my own purpose, I am somehow violating the rights of other people is absurd to me. He is denying me the right to my own ideas, and the ability to provide for my family.
Stallman is entitled to his beliefs, to communicate them, and to try to get them adopted. But I’m not surprised he feels pain trying to do this, personally I reject his beliefs. In my mind, I have fewer rights under his world view – I have no right to my own invention, I do not own any software I write, it must be released to the world. He is at least consistent in his beliefs, he feels the same way about music and movies and the companies that try to claim or exert ownership of what they create.
I do think the GPL license, conceptually, it a thing of elegance and beauty. Simple, concise, and powerful. I only have one problem with it – it is self-limiting. It limits usage to those who feel and think exactly one way. It also limits the ability of the established, mainstream market to participate.
It is the same with the Free Software Foundation. They don’t believe in intellectual property. I agree that the software patent system is badly flawed, but I still believe in ownership. But if you don’t comply 100% with their ideals, you are a bad guy. 90% is not good enough. There are no alliances based on common enemies. The FSF has no compelling, or viable, vision of the future that I can see. There is no plan. Just a skirmish war against a unbelieving world. Stallman even objects to the views and actions of Linux’s Linus Torvalds. Meanwhile the markets move on, alliances form, and invention continues. Everything evolves. The world evolves gradually. Unfortunately, Stallman’s views are more creationist – you cannot evolve to them, you will be shut out, berated, and branded as evil until you are 100% there.
Maybe at some time in the future Stallman’s views will prevail. Maybe the world will be a better place when that happens. I just don’t see it happening any time soon.
As can be seen from Levy’s article there is a divide between the hackers: those who view commerce as bad, and those who try to get their ideas adopted widely. Without commerce the world would be one big feudal system – the dark ages again. To reject commerce is to reject the world as it really is. That’s fine for an individual to do, it’s their choice, but to expect the rest of us to do us so in order to adopt or use their ideas is unrealistic.
When we look back in another 25 years at the effect the hackers have had on the world, who’s efforts will have had the biggest impact? The anti-commerce hackers, or the others?
The messages I get from the Windows 7 TV ads are
People who know nothing about computers had to tell Microsoft why Vista sucked so much.
Trust us at Microsoft, we don’t know what we’re doing
We’ll say how bad we are, so you don’t have to
Mat Aslett at the 451 Group has weighed in on the recent postings and comments between myself and the folks at Nuxeo, in a post – Let he who is without proprietary features cast the first stone
Matt raises some good points in a well balanced approach. Matt comes down on the side of Nuxeo not being open core, but not pure play either:
My own feeling is that Nuxeo’s approach is not open core, since the original definition of open core concerned proprietary products. However, the existence of Nuxeo Studio means that Nuxeo is clearly not 100% open source.
But he doesn’t have a term for what their model is, and is proposing adding a new one to their list:
For that reason, I have come to believe that we need to add a new revenue trigger category to our open source business strategy model, that makes a clear distinction between support subscriptions for 100% open source code, and value-add subscriptions that offer additional hosted services.
As yet, no-one is able to answer my question – if Pentaho stopped offering on-premise deployments and only provided a SaaS offering, would we no longer be open core? Maybe the answer is this new category that Matt mentions.
It looks to me like this model is close to an open core model – just hosted, not on-premise. But the term ‘open core’ does not specify how the non-open parts are deployed, just that some code is open source and some is not. So should we just create sub-categories of open core?
- Open core (on-premise), and open core (hosted)
- Local open core, and remote open core?
Runar Lie, founder of Office123, posted his list of the top 10 open source applications for businesses. Nice to see Pentaho on the list.
Red Hat CEO Jim Whitehurst, wrote recently about why car manufacturers should look into open source – Why Toyota Should Go Open Source
He has some valid points. We have a Bose system at home that about once a month stops responding to the remote and needs rebooting. If I add up all the time I’ve spent doing that (and talking family members through it over the phone), I’d gladly take a look at the source code and try to fix it.
When it comes to cars, however, I really like the approach taken by Local Motors. Their design and manufacturing process is very open:
- They have open competitions to select the car design. At this stage it is mainly concept art done by graphic artists and 3-D modelers. The range and imagination of the designs is very cool. The amount of work put in by the contributors is amazing. See the design wall.
- They have a voting process to select the designs that will be made into cars
- They have an open build process and CAD files are available – Build Process
- They have an open manufacturing process where you can participate in the building of your car – Build Experience
How cool is that?
Now they just need an embedded Linux-based OS for all the electronics that we can contribute to. Maybe some kind of serial bus that can be used universally by all the sensors and switches.
One of the participants in the latest round of open core bashing was Eric Barroca, CEO of Nuxeo. He chimed in with over 2000 words on why the open core model is ‘fundamentally flawed’ (his words). You can read his post if you click here.
In this post I am not trying to attack him or his company. But I think his post highlights how confusing and subjective this issue is.
Here are a few quotes from his post:
I’m deeply convinced that “open core” is fundamentally flawed
I would add that a subscription-based model (where subscription is for maintenance and support services) is superior (to an open core model)
I believe in clear business models. Successful companies use clear business models because that’s what enable trust from customers. Open core is not one of them: there is no clear line between open source / proprietary, neither serious justification for the customer.
That’s why we – at Nuxeo - won’t use the open core model even if it could increase short term revenue.
So Eric is fairly clear and outspoken about his view of the open core model. The following are snippets from the comments on the blog post
- James Dixon – Maybe I have this wrong, but doesn’t Nuxeo Connect include ‘Premium Tools’?
- Eric Barroca – Nuxeo Connect include SaaS. Not tools / code / additional features to apply on the open source software. Pretty much like RedHat for its OS or for JBoss. So, no, we’re not an open core company.
- Rahul Sundaram – So let’s keep this simple: Is any part of Nuxeo that you sell proprietary code is it all 100% free and open source? Otherwise, the comparison to Red Hat model is invalid.
- Eric Barroca – 100% of the code we sell / you install and run is LGPL. Period. So I think the comparison is very valid. :-)
Nuxeo Studio, a newly released configuration and customization environment for Nuxeo open source ECM, is available as a value-added component of the Nuxeo Connect subscription service.
Customers subscribing to Nuxeo Connect are assured of service level agreements for problem resolution, access to certified patches, value-added design, configuration, monitoring and management tools, as well as product updates in accordance with our product lifecycle policy.
When Eric says ‘100% of the code we sell / you install and run is LGPL’, I guess that’s true. But it seems there is code they sell (via SaaS), that is not installed, and is not open.
- Customers of Nuxeo have access to value-added design, configuration, monitoring and management tools. These features are not open source, they are proprietary, and only available to customers. These features are available via a SaaS offering.
- Customers of Pentaho have access to value-added design, configuration, monitoring and management tools. These features are not open source, they are proprietary, and only available to customers. These features are available on-premise.
As I understand it, having proprietary extensions to open source software – no matter how many extensions, or how many customers use them, or how they are delivered – makes you an open core company. As far as I see it the only difference between Nuxeo’s model and Pentaho’s model is that Nuxeo’s proprietary extensions are only available as SaaS, whereas Pentaho’s are available on premise.
If Pentaho stopped offering on-premise installations, and only offered SaaS, would Pentaho suddenly no longer be open core?
Do Nuxeo and Pentaho have different models, one open core, one not? If so, how come?
If Nuxeo and Pentaho have the same model, what model is it? Open core or what?
Who’s more confused, me or Eric?