Monday Sep 01, 2008

Happy Birthday To GNU

Stephen Fry

Today marks the start of celebrations for the 25th anniversary of GNU. It took a long time to get a working operating system, and a new paradigm to get it adopted, but the tenacious idealism of RMS has bourne fruit.

Like Barton, I am especially delighted that the Java platform has gone from being a case study in closedness to being the chosen technology for hosting the birthday video by polymathic British actor Stephen Fry.

Monday Jun 16, 2008

Firefox World Record

Download Day - English

The clock is ticking until 19:00 European time today when Firefox 3 will be released. I'm planning on participating in the Software Download World Record attempt that the Firefox community is planning. I was going to get Firefox 3 anyway, but that's an added incentive to grab it today since I've never been part of a successful world record attempt before. Join me!

Sunday Mar 23, 2008

Why Adoption-Led Is Not Trialware

In response to my posting Adoption-led is Not Shareware, IBM's Savio Rodrigues continues the exploration of the adoption-led concept in another blog entry. As an explanation of his view that "adoption-led" is a synonym for "shareware/trialware" he has modified my original explanation to read:

In this approach, developers select from available Shareware or Trialware and try the software that fits best in their proposed application. They develop prototypes, switch packages as they find benefits and problems and finally create a deployable solution to their business problem. At that final point, assuming the application is sufficiently critical to the business to make it worthwhile to do so, they seek out the creator of the Shareware or Trialware to provide support, services (like defect resolution) and more. Adoption-led users are not all customers; they only become so when they find a vendor with value to offer.

Savio is correct to note, as Zack does, that Free software is easier to evaluate. The trap Savio is falling into here is to go no further and to be focus only on the procurement of the software. That's a natural reaction given the dominance of procurement-driven thinking, but it hides the natural progression of software in an adoption-led context. What matters is not the freedom to procure the software, but the freedom to adopt it. The two are only the same in a procurement-driven world.

Freedom to Use

No Fun

With trialware/shareware, one is only free to use the software for evaluation purposes. Once one switches to production, one is compelled to pay for a license to the software in order to take it into production. This causes significant problems in an adoption-led context:

  • It frequently means that the features needed for production are not present in the software in use.
  • It erects a barrier between evaluation and deployment that most likely involves a switch of versions, resulting in a need for re-evaluation. For example, if the evaluation was on Fedora, to go to supported production the system will need to move to RHEL, since Red Hat do not support Fedora (as far as I know).
  • It erects a payment barrier in the adoption process, one which is located at an arbitrary point known only to the vendor's legal team. Does one have to pay between evaluation and development? Between development and pilot? Between pilot and initial roll-out? Between initial roll-out and full production? Who knows?

True Free software (software with the freedoms left in) doesn't have these problems. There are no artificial barriers or toll-booths on the path to deployment, and adoption is a process that can be conducted under the control of the adoption team. As they progress they can decide at each stage whether to go buy a subscription from a vendor to meet their sustaining needs, or whether to engage experts from the open source community around the code to do it, or whether to just absorb the risk themselves.

This is the key to the adoption-led approach. It's about adoption of software, not about the procurement of software. It puts control in the hands of the adoption team. It's about freedom - freedom from vendor control, freedom to use the software how you want.

[Previous: Adoption-led as a Force of Nature | Next: An Adoption-Led Business Model]

Sunday Mar 16, 2008

Software Freedom: More Than Copyright

New Forest Reflection

I was surprised last week to see a posting from Michael Tiemann, the President of the Open Source Initiative and a VP at Red Hat. Any posting with a subject of line of "Simon Phipps Was Right" is bound to catch my eye, but this one was especially unexpected because in the original discussion I had thought that Michael was largely right! Michael's posting graciously said:

Simon, I'm beginning to think that you were right and I was wrong. You said a standard's process is a crucial aspect of the standard's product, and a process that is not open cannot be trusted to produce a product that can be considered open. I maintained that I had seen and used many wonderful standards that took absolutely zero input from me, and therefore I didn't see my participation as a necessary prerequisite for assuring quality in the future. I believed that no matter what the process, a standard should be judged by the product. Watching the fallout settle from the [ISO ballot resolution meeting] in Geneva, I'm beginning to think that you were right and I was wrong.

I've been thinking about the posting for a week or so now and I've tried to respond thoughtfully. Here is the response I sent to Michael (still awaiting moderation):

Thank-you, Michael - it's not often I see a posting like this. Actually, when we spoke about this at OSCON I found I agreed with many of your arguments, even if that doesn't show in the on-list discussion. The problem is that standards are orthogonal to open source, and attempting to define them in a way that promotes and protects software freedom may be impossible. It's been said that when we create any system we create the game that plays it. The standards system is fully mature and as such is fully gamed, as the DIS29500 debacle you reference is proving.

Maybe a more productive approach going forward is to try to do for the other kinds of so-called intellectual property what the Open Source Definition (OSD) currently does for copyright licensing. Perhaps we need to rename OSD to "Open Source Copyright Definition" and then work on an "Open Source Patent Definition", so that we can avoid the kind of entrapment that software patents can threaten? And as you know I am convinced we need an "Open Source Trademark Definition" to help us as a community of communities to avoid the IceWeasel problem.

If these are interesting, I'd be pleased to spend time exploring them together. Let me know.

New Definitions

The current Open Source Definition doesn't actually define Open Source - rather, it defines a subset of the requirements that protect software freedom, in this case the copyright license. I actually think renaming it ("Open Source Copyright Definition"?) would be good since there's more to Open Source than just the copyright license. I then suggest we explore creating an "Open Source Patent Definition" and an "Open Source Trademark Definition".

What would be in these two new definitions? Both would need to define what promotes software freedom and how it can be protected. Both would need to be pragmatically principled.

  • An Open Source Patent Definition would do for patents what the OS(C)D does for copyrights. I've posted a lot on this subject before, notably in Protecting Developers from Patents and Ten Reasons The World Needs Patent Covenants, so I'd go mining there for my contributions to the discussion. But it may also be that in addition there needs to be a call for patent law reform, maybe as I outlined in Seven Patent Reforms While We Wait For Nirvana.
  • When it comes to an Open Source Trademark Definition, we would need to similarly define the signs that a developer or user needs to know whether software freedom is being promoted in a trademark policy. I've not written about this yet, but I do believe we need to collectively understand the bounds trademark law places on people who have responsibility for trademarks (read: all developers and open source communities as well as all vendors). We then need to construct a path that promotes software freedom without placing impossible demands on trademark owners to behave in ways that are contrary to their responsibilities.

This is not easy stuff. But I do believe that certain recent events between the open and proprietary software worlds mean that it's time for software freedom fighters to get together and work on these things. I'm ready to work on it. What do you say, Michael?

Thursday Mar 13, 2008

Adoption-Led as a Force of Nature

Crater Lake Sunrise

In discussing how the software market is increasingly an adoption-led one, a frequent point of departure is to look at ways in which existing software companies are pursuing open source models.

Centralised to Distributed

But the idea that the adoption-led model is a go-to-market strategy created by software vendors is wrong. Although as Zack observes it has become a successful driver for some companies, it is fundamentally a consequence of a set of social changes which in turn are the consequence of the pervasive nature of the Internet. No amount of debate for and against an adoption-led business model will ever change the fact that the market is moving that way.

The first mechanised communications - dating back to before the Industrial Revolution - helped to create a hub-and-spoke social topology. The ability to communicate to a large number of people was necessarily centralised and recognition of authorship became more important. Interestingly, this is when copyright law first emerged. And as more communications became industrialised, so did society become more centralised. Author at the hub, readers at the spokes, suppliers at the hub, customers at the spokes, government at the hub, citizens at the spokes. The Web is changing that. The topology is changing to a mesh that even crosses cultures and borders. Peer-to-peer is the new order.

While the internet has existed for a relatively long time in technology terms, the Web as an application has driven it to ubiquity in a very short time. And this is what gave the equally long-standing Free software movement the vehicle it needed to influence the mainstream. The two together have placed a growing wealth of software within reach of every sufficiently skilled developer, giving them the freedom to use it however they wish. As Stormy points out, they can now bypass the whole existing system.

Demands a Response

It's this sudden wealth of choice which created the adoption-led movement that I described before. It didn't - and doesn't - need vendors to happen. Rather, it demands a response from vendors. Some try to ignore or to discredit it. Some pay lip-service to it, using its fruits but shunning true participation. And some embrace it, employing people to work within open source communities. Each of these approaches has business models associated. None of these approaches have themselves caused the adoption-led market to spring into existence.

Now, although this movement did not need vendors to make it happen, vendors have the opportunity to help it into the mainstream for business usage. As I suggested in my response to Savio, there's a lot of value that a vendor can add:

The model assumes that enterprise users will want the value-added content of a "subscription" or "enterprise version". Value-add can include patch management, performance tuning, additional utilities and more. Corporate governance regulations may make enterprises using software for a mission-critical purpose require a service contract, or seek a warranty for their software infrastructure. Those who are embedding software in their own product may require indemnification. Finally, many businesses are reluctant (for whatever reason) to use open source licenses and so want commercial licenses for their production systems.

To deliver on that value, it's my belief that the biggest opportunities lie beyond the mere aggregation of the work of others (although that does seem to be a viable option for some). I believe that through influencing the direction of a project, through employing committers to an open source code base, by creating new code, by being a responsible community steward and by bringing leadership to the challenges open source faces, a software vendor can take full advantage of the opportunities the adoption-led market presents. And I believe that success will be proportional to the contribution made. No free lunches, at least for those wanting job security.

Led, not Driven

This is the response of the software industry to the mesh topology. It's one where copyright, through open source licenses, is used to foster creativity, rather than to restrict access (and the inverse for patents - a reversal worth discussing elsewhere). We are moving from the "procurement-driven market" to the "adoption-led market". One is driven by vendors. The other is led by deployers and developers. That's the key, and I think other industries should examine with interest the lead that the software market is providing, since I expect the phenomenon to spread beyond software.

[Previous: Adoption-led is not Shareware | Next: Why Adoption-led Is Not Trialware]

Open Source for Sovereignty

Parascending in Sydney Harbour

I was interested to see news from the European Commission where they announce a new policy to more frequently use open source software in the administration of the European Union. They say:

For all new development, where deployment and usage is foreseen by parties outside of the Commission Infrastructure, Open Source Software will be the preferred development and deployment platform.

It's not just European government that's opting for open source. Today the NSA (the super-secret spy agency in the US) have announced they are joining in with OpenSolaris. Barton has an interview that explores this more. I think we'll see more and more government engagement as the adoption-led market takes hold.

Using Free software from open source communities makes perfect sense for governments, and not just for the obvious reasons of up-front savings on license fees. As I heard said on behalf of the Brazilian government, open source is a matter of sovereignty. When a government decides to use closed software, they are guaranteeing that they will be sending money out of the local economy. The degree of expatriation depends on the actual system they've chosen. In the worst case, all the money goes to the US, all the resulting assets belong to someone else and all the ongoing service and support costs pay for the development of skills abroad.

By contrast, using Free software has no licensing costs. Any extra programming results in an asset shared by an open source community. All service and support can be handled locally, growing the skill-base and economy. What could be a smarter way for a government to obtain the essential infrastructure it needs and develop the local economy at the same time?

Sunday Mar 02, 2008

The Adoption-Led Market

Fruit Bats, Asutralia

I've previously spoken of payment at the time of deployment rather than at the time of selection - Software Market 3.0 - as the defining characteristic of open source business models. As I've spoken about this in all sorts of contexts, it's become apparent that this is just an aspect of a deeper trend, which in some ways relates to Greg's red-shift/blue-shift idea.

Traditionally, the process of acquiring software has involved a request for proposal from vendors against a customer specification. Vendors then make proposals, submit prototypes, contend for business. In smaller bids, an evaluation team considers trial versions, makes evaluations, makes proposals to management. Eventually software is selected and paid for. At that point, adoption can begin. Every user of software in this model is also a customer. Software selection is something of a matter of faith in the procurement-driven market.

Defining Adoption-Led

The switch to a mesh topology for society1 has led to easy access for everyone to Free software created by open source communities. The result is an emerging approach which is rapidly spreading for smaller software projects and in my view is the future of all software acquisition. The emerging approach is an adoption-led market.

In this approach, developers select from available Free software and try the software that fits best in their proposed application. They develop prototypes, switch packages as they find benefits and problems and finally create a deployable solution to their business problem. At that final point, assuming the application is sufficiently critical to the business to make it worthwhile to do so, they seek out vendors to provide support, services (like defect resolution) and more. Adoption-led users are not all customers; they only become so when they find a vendor with value to offer.


Written down like that, it seems pretty obvious, but having a name for it – an adoption-led market – has really helped pull together explanations and guide strategy. For example:

  • In a procurement-driven market you need to go out and sell and have staff to handle the sales process, but in an adoption-led market you need to participate in communities so you can help users become customers.

  • In a procurement-led market you need shiny features and great demos, whereas in an adoption-led market you need software that is alive, evolving and responsive to feedback.

  • In an adoption-led market you need support for older hardware and platforms because adopters will use what works on what they already have.

  • Adoption-led users self-support in the community until they deploy (and maybe afterwards if the project is still “beta”) so withholding all support as a paid service can be counter-productive.

Naturally now I have a new hammer everything looks like a nail, so expect to hear more of this if you talk with me!

[Next: Adoption-led is not shareware]

1. Described in more detail in Adoption-led as a force of nature

Tuesday Feb 26, 2008

MySQL Joins Sun Today

As you have no doubt noticed by now, the Sun acquisition of MySQL completed today and MySQL is now a part of Sun Microsystems - Jonathan has a welcome letter. I actually knew Mårten Mikos, and David Axmark already from our shared activities campaigning against the European patent directive a while back, but over the last two months I have been reacquainted with them as well as meeting a huge number of MySQL staff. They are a great company with a strong engineering and community culture and I think they will be a great fit in Sun.

There's still a whole lot more work to do though. One important task is to introduce everyone at MySQL - staff, community and customers - to everyone at Sun, and a key part of that is the World Tour that's getting started - Kaj Arnö has more. You can also hear Barton George from the Sun Open Source group interviewing Mårten and Zack Urlocker.

Saturday Feb 16, 2008

FOSDEM Approaches

I’m going to FOSDEM, the Free and Open Source Software Developers’ European Meeting

Want to meet up at FOSDEM next weekend in Brussels? I'll be at the Java Libre dinner on Saturday evening, or we can exchange electronic tags to try to meet elsewhere. Leave a message with a valid e-mail (it's not displayed here). Full details of what the Sun team will be up to are on our wiki.

Tuesday Jan 29, 2008

How To Get $1m From Sun

You'll remember when I was in India I announced a $1m programme from Sun to recognise contributors to a selection of Free/open source communities (and gave a few more details afterwards). In the interim, each of those communities has formed a team to work out how they would like to direct the $175,000 Sun has allocated to them, and I'm delighted to announce that the six programmes are now live. Full details in the press release which links to each of the community sites.

Tuesday Jan 15, 2008

Swimming With Dolphins

You may have noticed it's been a bit quiet on my blog for the last few weeks - just the tumbleweed rolling through lists of links. While part of that was due to the Christmas break, it's also because I've been working on a team at Sun on a very significant project that we've been keeping a close secret as a matter of legal necessity.

But today the covers are off and I can disclose the nature of that project. This morning at 8am EST, Sun announced that it had entered into a definitive agreement to acquire MySQL AB for a consideration approaching $1bn, subject to all the usual approvals. Jonathan has a long discussion of the news. In support of this move, Sun will be unveiling new global support offerings into the MySQL marketplace. We'll be investing in both the community, and the marketplace - to accelerate the industry's phase change away from proprietary technology to the new world of open web platforms.

I've been part of the due diligence team and having seen all the data, I think this move is a great step for Sun, MySQL and the FOSS communities. For Sun, because it means we now have the ability to offer support on a full web operating system - OpenSolaris, Glassfish, MySQL, NetBeans and much more. For MySQL, because it means the next phase of growth can ride the proven global organisation Sun provides to delver support and other value to customers.

For the global community of Free and open source communities because MySQL is in safe hands. Sun is committed to preserving all the strengths we've come to expect from MySQL - broad platform support, high quality engineering, frank and transparent community engagement - all under the GPL. This joining of strengths is additive, not subtractive. Just as with, Sun is aware that MySQL is a crucial FOSS community asset and we intend to be a good steward, balancing community and commercial considerations.

I already know plenty of the folk over there but I'm looking forward to welcoming the MySQL staff (who are just finding out about this now at their meeting in Orlando - Kaj has the news) as colleagues. The next wave of Free/open source is starting, and I've dolphins to swim with!

Tuesday Dec 11, 2007

About Sun's Million Dollar Grants for FOSS

Tiger in Bannerghatta

I just left India after speaking at FOSS.IN where I announced details of the Sun Open Source Community Innovation Awards in the context of a talk [20Mb PDF] about the challenges the Free/open source community-of-communities faces from success and growth. There's no question that India is a 'waking tiger'. The energy and enthusiasm I have found here has been without peer on my travels - apart, perhaps, from that of Brazil.

I have been considering with interest the reaction to my posting last week foreshadowing the awards. The awards were widely applauded, although as I'd expected there was also no shortage of people wanting to attack the program. I feel that some of the adverse reactions illustrate only the biases of their authors, and others were the result of of incomplete information. So I'd like to clarify a few points.

Programme Structure

The press release for the Open Source Community Innovation Awards uses the word "prize". As anyone that has tried to construct a similar program will know that to pay grants in this manner, laws surrounding prize-bearing competitions in the USA and elsewhere may be applicable. We've actually made no detailed decisions about how the $1 million US fund will be distributed beyond believing it will be divided equally between the communities.

Instead, the six communities involved will each devise schemes that fit their members and bring the proposals to us, for us to knock them into a shape that complies with the laws in the places the communities want to address. We are wanting to recognise and reward innovation, which we fully expect to come mainly from existing community members including the many already employed to work on software (though not Sun employees since Sun is the sponsor of the awards). It's about sharing the wealth.

Yes, this may incent some people to join communities, but frankly I'm well aware that developers are motivated more by their own goals with the software than by this sort of award. And I'm therefore astonished to come across the notion that Sun is looking to attract "cheap labour" with these awards. In my opinion, people that subscribe to that kind of view of open source fundamentally don't understand what community development is all about - they're either confused, or trying to confuse somebody else, to borrow a phrase.

I hope that the communities will devise a variety of programmes that include "innovator of the year"-type awards, "feature bounty" schemes, "internship"-style project sponsorship and other styles. The awards will be annual and if they work out well I hope we'll be able to expand their reach. For the first year, we've picked a small group of communities that Sun knows well since the legal and administrative details turn out to be pretty complicated. However, by their very nature, open source communities are networks of interests. OpenSolaris, for example, includes work in a range of communities including GNOME,, Mozilla and many others. The 2008/9 scheme will learn from the experience of the 2007/8 scheme and hopefully be even better - we all have to start somewhere!

Global Reach

The communities are welcome - encouraged - to apply the fund globally. The fact I announced the scheme in India doesn't imply it's targeted only there. However, I have personally seen that places like India, China, Brazil and Malaysia (which I also visited on this trip) have an enthusiasm and energy about FOSS that is raw and fresh, and I fully expect so see the people of this region well represented on the list of beneficiaries next year.

The next step will be for the communities involved to form proposals for their individual use of the $175k US or so they will be distributing. If you're a member of one of those communities, expect to see details of how this will be done coming from your leaders soon. I'm expecting to be able to join them in announcing the schemes early next year. If you've other questions, please ask in the comments and I'll try to answer them.

This scheme has proved surprisingly challenging to put together, and only goes so far in rewarding FOSS developers. I still think the best way to do that is to hire them, and indeed Sun does that to the tune of more than $200m US each year. But this scheme (and the others like it) is a useful addition and I hope the innovative features that result from it will greatly enrich the world of Free software and the open source communities that develop it.

Update Jan 2008: Details of the individual programmes are now available.

Monday Dec 03, 2007

Getting Paid to Develop

Beaver photo (c) S Phipps

We've got an exciting development bubbling that I hope to be able to announce in full detail at FOSS.IN in Bangalore on Friday when I speak there.

Just to give you a glimpse of what's happening, Sun will be announcing a multi-year award program in support of fostering innovation and advancing open source within our open source communities. We'll be providing a substantial prize purse and working with the communities involved to develop the approach that works best.

I'm announcing it in India because that's where I expect the greatest open source community growth to come from in the near future - the FOSS.IN programme committee relaunched their CfP a while back with this in mind. If we can play a part in catalyzing the emergence of India as a key international open source power-house, the effect on the software industry will be huge. Not to exclude others in the region of course, so much is going on there.

This year's participants include OpenSolaris, GlassFish, OpenJDK, OpenSPARC, NetBeans, and This is a great opportunity for members of these open source communities to take their passion and creativity and push the innovation boundaries - and get paid in the process!

Update: I've added more detail as well as information responding to questions in the comments here to a new posting.

Update Jan 2008: Details of the individual programmes are now available.

Wednesday Sep 05, 2007

The Lure

I see Miguel is expecting flak for his initiative to implement Silverlight on GNU/Linux, and I'm sure he'll get it. The thing that caught my eye, however, was what terms I was asked to agree to if I as much as give Silverlight a try on any other platform in the ecosystem Miguel is helping create. Just take a look at the license agreement you're assumed to agree to if you so much as click the "Get Silverlight" button (yes, your acceptance is there in 4-point text in the Get... graphic). You will be agreeing you will not:

  • work around any technical limitations in the software;
  • reverse engineer, decompile or disassemble the software, except and only to the extent that applicable law expressly permits, despite this limitation;
  • publish the software for others to copy;
  • rent, lease or lend the software; or
  • transfer the software or this agreement to any third party.

In addition to that, you are agreeing:

  • that the limit of Microsoft's liability in any matter (including "internet services") is $5;
  • that Microsoft can gather information about your computer and internet connection;
  • that they can automatically modify the software.

Update:As Miguel repeatedly points out, if you're smart enough to actually read the license you'll discover you can opt out of these last two defaults. Most people won't.

If you're a business, you're also not covered by the MPEG patent licenses, and you may be agreeing to waive some of your other contract terms if you're a Microsoft competitor.

Now, I'm not a lawyer but to my eyes, all those are terms that are incredibly hostile to the spirit and practice of Free software. Miguel is encouraging you to surrender your freedoms if you're using the technology he promotes anywhere but the operating system he is working on. He's the lure for someone else's trap.

That's also why I don't yet share Tim O'Reilly's enthusiasm for Microsoft's apparent epiphany. I realise it will take them a long time to modify their historic behaviours, and I would welcome their decision to promote software freedom. Their new projects, however, should not be actively undermining our freedoms like this, and I can only conclude that regardless of any lip-service to "open source", software freedom is not yet their goal. That's the issue, Tim.

Update: Luis has a great photo and also a great point about patent protection. Also, to be clear, Moonlight is under a Free license, not under the EULA I'm commenting on above. My issue is rather that it acts a the lure to an architecture designed at best with no regard to software freedom and at worst withthe intent of removing it.


Thoughts and pointers on digital freedoms and technology markets. With a few photos too.


« July 2016