Arjen's blog

Carribean opensource

It's been quiet on this blog. A long vacation that was not really a vacation, more on that later.

Last week I was visiting the Dutch Caribbean by invitation of the local government to do the opening keynote (ODF 1 2) on their conference on open standards and opensource. Curacao, one of the islands of the Dutch Antillies is about to become fully independent nation state and that means a lot of re-design of the local IT systems of the government and public sector. The government is determined to maximise the opportunities offered by open standards and opensource software to move the new governemnt, the local educational system and economy forward. For education an OLPC project is being considered because 3 PCs per school is not the way into the 21st century. Hopefully the new Internet Exchange (based on the Dutch one) will bring down the cost of bandwith so that all those OLPC's will be able to go online.

All of this will not be easy to achieve. Curacao is a very small entity to function as an independent nation and do everything themselves. There is a great need for expert knowledge and training to bring local IT-staff and administrators up to speed and the available budgets for this are very limited.Curacao spends about 20 million pounds per year on proprietary software licenses (mostly in government and other public sector), this amounts to about half a month's wages per citizen. If this can be reduced by 30-50% the budget required to make the desired changes is available (assuming the total IT-budgets can be kept at the same level for the time being).

Because this visit also included several meetings with local dignitaries and media appearances I was invited to stay for a whole week in a beautiful apartment west of Willemstad. Many thanks to Ace Suares who has been working for open-IT on his island for many years. He was the driving force behind the whole conference and a wonderful host.

The Last HOPE

Just finished my talk at the Hope conference about the national Dutch open standards / opensource policyplan. Hope stands for Hackers On Planet Earth and has been a bi-annual event since 1994. HOPE features presentations on a range of (IT) security related topics but also hosts all kinds of techno-creative workshops and lectures on the social relevance of technology. For a full listing see the conference site. Video's of the talks will become available in the future.

Because the mid-town Manhattan hotel just across from Penn station in New York may be demolished next year it is unclear if there will be (a) hope after this one. Because of this, and a certain facination with morbid themes, the conference organisers are calling this 'The Last Hope', refering both to their conference and the state of (digital) civil liberties in the US.

Cnet has the first coverage on the event.

My slides are here in ODF and PDF. For more background links and video see the 24C3 page (Dec 2007 talk in Berlin) and my blogpost about LinuxTag (May 2008 Berlin).

LinuxTag presentation

On Saturday May 31st Gendo presented its experience on changing national policy on open standards and opensource at LinuxTag 2008. The link to the talk description is here and the slides are here in ODF and PDF. LinuxTag is Europe's largest opensource conference with over 11.000 participants from 31 countries. Doing a talk there was a lot of fun and I'm looking forward to next year.

Linux Magazine Germany has a writeup of the talk here. The presentation was not recorded but an earlier (similar) presentation was taped last december at the CCC congess, also in Berlin.

Conservatives want opensource

Conservative party leader David Cameron held a speech at the National Endowment for Science, Technology and the Arts on April 3rd. Aside from the usual criticism (not all undeserved) at the current government he talked about the network society and bottom-up collaboration enable by ubiquitous access to IT. He (or his staffer) had obviously read "Mavericks at Work: Why the Most Original Minds in Business Win" by William C. Taylor and Polly G. Labarre. And a good thing too since it is filled with good examples of how to innovate by sticking your neck out and doing things differently.

About three quarters through the speech Cameron very explicitly stated that his government, if elected, would actively seek a greater use of opensource software and mandate the use of open standards for public IT systems. According to him this will enable faster innovation in the public sector while avoiding repeats of the very expensive NHS disaster.

When things like this are being stated by a conservative the idea has arrived. The other political parties will have to follow suit or be left in the dust by the conservatives. I think these are really good ideas and I know they are practical because I advised the Dutch government on a policy plan (full English translation here) last year that was adopted by parliament last year and is being implemented now. Since the publication of the plan that has been a remarkable change in the relationships between many government institutions and their usual software suppliers. They customer is re-taking power by forming interest groups and re-defining what IT-vendors should actually deliver (instead of just waiting what vendors want to sell them). Norway has a comparable policy and France, Germany Italy and Spain are taking concrete steps too. Below an example of the effects such a policy has had on the quality of healthcare and education in the south of Spain:

More about my work on open standards and opensource here.

A tale of two elections

There were two elections on my news radar the last two weeks. One in an African country now rapidly sinking into an economic crash and possible civil war, the other in the heart of the civilized west; the OOXML election by ISO standards body in Geneva. Aside from the locations and prices of hotels for journalists the differences were few. Bribery, fraud and intimidation were applied to achieve a specific outcome, never mind what the majority wanted. In the OOXML election even the standards organizations of proper democratic countries like Norway and Germany were unable to withstand the well-oiled lobby machine of the worlds most convicted software monopolist. In the Netherlands Microsoft was actually a member of the committee and prevented a committee consensus. So instead of voting against the standards, as 21 out of 22 members wanted, the Netherlands voted 'abstain'. The list of irregularities is endless but try some of the links above to get an idea.

To quote an ISO insider: "“OOXML’s BRM process is irretrievably broken; complete, utter, unadulterated bullshit”" (Tim Bray). The ISO process is dead and that is bad. Even though ISO was far from perfect it's open process did help many nations to agree on thousands of standards on anything from steel manufacturing to surgical gloves. These kinds of standards are crucial for a globalized high tech society.

It would seem that those in favor of the OOXML standard (Microsoft and a few of it's partners) did not have so much confidence that their 6000-page specification would make it to ISO standard status that they trusted a transparent en democratic process to take it's course. There way of treating the process was much like an African dictator with a chest full of medals. Utterly lacking in legitimacy but high on power as a consequence of an unlimited bribery budget and a willingness to bully anyone in the way of their goal. If they were not so much trouble you'd almost feel sorry for them.

Microsoft is of the opinion that everything went fine. Neelie Kroes and the EU have a different view of the matter. I feel another mega-fine around the corner. Not that any of this matters a lot to the fastest growing parts of the world. Brazil, China, India, Iran voted no. They represent a big chunk of the global population (as opposed to yes-voters such as Malta). If I were an applications vendor I'd make sure I stay ODF compatible. Several countries such as the Netherlands and Norway (that is now appealing their stolen vote) have already formally chosen ODF as the document standards for government. Others are looking to follow this lead. It's not entirely clear what Microsoft has achieved by gaining the ISO label. The governments that already choose ODF are not going back and Microsofts well documented behavior is not going to make then many new friends.

The EU appear to have been investigating the OOXML ISO certification process for some time now. Documentation from countries with where very clear breaches of protocol occured will strengthen the case for further antitrust investigations and possible additional fines or other measures. The victory of OOXML may turn out to be very costly indeed. And any credibility of OOXML as an 'open' standard has already been severely damaged by the deluge of reports about bribes, fraud and other forms of misconduct by Microsoft over the last year. A Pyhrric victory if here ever was one.

Desktop mono-culture for the Dutch

The Dutch national government is proceeding with its new implementation of 'GOUD' (Gold), its new plan for a 'standardized desktop' for ministries and other governmental institutions. After the much publicized new Dutch National IT-policy of last autumn this might sound wonderful. It's not.

The problem with the 'standardized desktop' is that it uses very few real standards. Instead it uses products and pretends they are standards. The difference between 'standards' and the idea of 'standards products' is widely misunderstood, often confused and this is where projects such as 'GOUD' go off in troublesome directions.

Standards are things like Open Document Format, HTML (the programming language of web pages), LDAP (a protocol for defining access-control on a network) or TCP/IP (the low-level network protocol of the Internet). What all these things share is that they are not owned by a single entity. They are a formalized form of knowledge about how to achieve certain technical objectives. This knowledge can be used by programmers to build tools that actually do these things for us. This means that there will be many different, competing, implementations of the standards that can inter-operate. This diversity fosters a working market and ups the quality while lowering the price (often down to zero).

'Standard products' on the other hand are, by definition, products and generally owned by private corporations with a profit motive. Standardizing on a product is essentially creating a mono-culture of one type of product, making yourself totally dependent on the vendor of that product. In the case of desktop computers this is extra problematic because there exists a de facto monopoly in this market. Creating a mono-culture in a monopolized market actually re-enforces the monopoly, the exact opposite of the aims of the adopted policy plan.

The Dutch government (or any other for that matter) should not work towards a 'standard desktop' but towards open desktop standards. This is precisely what the Dutch parliament unanimously requested from the government in 2002. Open standards, such as the ones mentioned above, allow computers of many different types and the people using them to communicate.

The Internet is a network of billions of computing devices made by thousands of different vendors and these can all communicate through open standards. Because of this I can e-mail other people without having to know anything about the computer they use. As long as we all agree to adhere to the standards information sharing is possible with a variety of systems and products. Open standards make this possible and should be the cornerstone of any governmental IT-architecture.

I've had many conversations with civil servants on this issue over the years and I conclude from these that the vast majority of them have no conceptual understanding of the difference between (the use of) standards and product mono-culture. In 2005 I was speaking at the Ministry of the Interior to a civil servant who enthusiastically talked about the 'standard' desktop they were developing for the entire government (350.000 desktop in the Netherlands). When I asked him what standards they were using he responded with a list of products: 'this version of an operating system, that version of a word processor', and so on. I tried not to cry. I really did.

In the 'Gold'-documentation (yes, the project name is a gift to punsters) the same way of thinking is pervasive; in order to collaborate, all civil servants need to use the exact same software products. I call bullshit. One of the primary recommendations of a counter-expertise report was to replace product-specific standards by open standards. This would enable interoperability with a variety of systems in the future and thus free the government of its dependence on a single vendor. This remains one of the primary stated goals of the national policy set by the government and applauded by parliament.

I can imagine how 15 years of vendor-lock-in-by-standards would lead people to believe computer systems can only share data if they are all supplied by the same vendor. This is however not a law of nature but a symptom of a badly functioning market. It is a problem to be solved by making a policy plan and sticking to it. The national action plan obliges the government to actively seek solutions to the problems of interoperability by the mandatory use of open standards.

In a letter to parliament the responsible minister for the 'Gold' plan claims two things:

1. With the new 'standard' desktop civil servants will be able to collaborate better. It is not clear why they cannot collaborate now. Don't they all have e-mail and a phone? What 'collaboration' and 'better' means in this context is never defined so it's kind of hard to verify this claim.

2. By implementing a mono-culture of Gold PC's the cost of systems maintenance and administration will be reduced to a level that will compensate the cost of procuring and implementing them in the first place. Not that this claim requires a detailed understanding of the existing level of cost.

MPs I work with will be seeking freedom-of-information acces to all relevant documents pertaining to this project from 2006 onwards. I'm looking forward to being proven wrong but I'm not holding my breath.

7-year olds get school laptops... in Peru

Wouldn't it be great for education if all school kids had their own laptops? Well, in Peru this is happening. And in Uruguay, Mongolia, Libya and Ethiopia. The laptops are designed and developed by a non-profit foundation called One Laptop Per Child (OLPC) started by Nicolas Negroponte, formerly of the MIT Medialab.

The OLPC is a technical miracle. It was in fact considered impossible by all the major commercial manufacturers and software builders like Intel, Dell, IBM and Microsoft. That did not hold back the OLPC team though and they build something not only affordable ($150-$180) but also rugged, maintainable and very, very eco-friendly. My Mac laptop eats 30-80 watts, the OLPC does 1-2 watt. That matters when your primary source of electricity is a $10 solar panel. If the world switched to OLPC's from their current systems the savings on power cost (even excluding the environmental cost!) would be sufficient to buy 50 million OLPC's per year for the third world.

On the software side the OLPC foundation chose to use open source software wherever possible, both to keep the cost to their clients as close to zero as possible but also to allow local communities to adapt the software to their needs, wishes, languages and cultural conventions. All the specifications of the laptop (both hard- and software are available for public scrutiny and improvement from the OLPC wiki. This guarantees the independence of nations basing their education on OLPC devices, they will remain in control over the content and methods of education because they have full control over the technology.

But the technical brilliance is not the most important part of the story. The way digital teaching aids allow for new ways of disseminating knowledge is. When every student has a networked digital device that can store and display digital information, schoolbooks can be fully digitized. That gives opportunities to free the ownership of the information in them from the classic model of big publishers that exert control over what goes in the books. A schoolbook is then just a part of a wiki-type system where teachers, experts and other interested parties can collaborate on educational materials that is then shared freely with the world. Some of these contributers may be financially rewarded for such work in the same way we now financially reward publishers. The output however would belong to all of us and we would not have to buy the same knowledge over and over again.I've written more about this here.

The interesting question for western countries is whether they will lag behind places like Peru and Mongolia in modern education. Western European countries typically spend 150-300 euros per child per year on schoolbooks so there's no question that the budget is there. It just needs to be spent slightly differently.

Much has been made of the fact that teachers would not be able to handle the technology. But teachers don't teach kids these laptops, they mostly teach each other. But don't take my word for it. Listen to the 9-year old expert:

Some interesting ideas from Robert X Cringley on the future of education if all kids have full access to online information resources. As often with these type of articles there are as many good ideas in the comments as in the article itself

Weekend in London

Easyjetted over to London on Friday after finishing my 'Train-the-Trainer' class. The students worked hard on writing responses to two articles in a Dutch E-zine on open source/open standards. While bored on the plane I added my own article (Dutch) to the discussion.

Now working on the Makewarshistory campaign website and talking to journalists on Monday about open source for governments, voting computers and the Mifare RFID hack. Trying to find some free time to spend doing as little as possible with my love Annie.