Open Standards and Open Source at Smartgov Live

OSC Presentation at SmartGov Live

On 15 June, Gerry Gavigan gave a speech at SmartGov Live, in the Session “Open Standards and Open Source Software: Crucial for Government IT in a Big Society”

The text of the speech is reproduced below.

I think it would be difficult to have any interest in today’s session and not be aware of the number of pronouncements made by this government and the last, regarding open source software. Apparently, they want to encourage more use of it. It might come a surprise that I think this is a bad idea.

As long as open source software is perceived as needing yet another government initiative, we are all losing out. So you might be pleased to know that, today, I will not be demanding that government uses more open source.

I’ll be talking about three things open standards and interoperability, agility and localism. And to conclude, I’ll explain why I think that if you sort all this out you don’t need to do anything about open source software

So then, open standards and interoperability

It seems peculiar that we even need to be talking about this: Industry gets standards, government gets standards so why don’t government IT customers get standards? Standards are so important that I can do no better than quote a government study published jointly with the BSI only 2 years ago:

Standardisation is a key factor in support of a number of government policies, including competitiveness, innovation and public procurement. Its importance is growing with the convergence of technologies and a growing knowledge economy. However, standards can be used to create barriers as well as to remove them. Government seeks to prevent this and promote the very beneficial effects of standardisation.

So really with all those warm words you’d think there was nothing more to do. And if you’d been following the debate in January 2011 you might have thought even those in government IT get it.

The definition on open standards contained in the OGC policy procurement note was a good one but it was undermined by two features:

  • the first that open standards were only to be used “wherever applicable” which includes nowhere
  • the second that even with open standards you need to separate programme features from the data to ensure interoperability

For engineers the second point is merely good practice.

Government IT types, however, so encumber their data files with code and features that it is impossible, for example, central design means that local authorities have to use Microsoft Office rather than one of the many alternatives.

As a member of the public you will find similar challenges including forms for completion or indeed if you want to watch democracy in action you will increasingly be invited to download Microsoft Silverlight. This includes an increasing number of local authority council meetings and even Parliament, despite the many technology neutral alternatives that could have been used.

If that were not enough only three months after the definition of open standards were published, it was diluted to being a draft – with a Cabinet Office spokeswoman quoted saying “it was not set in stone”

The dilution took place in a Cabinet Office survey on standards, published in March. So there was are really, despite all the government identified benefits of standardisation – government IT carries on getting it wrong.

And I think I understand the reason why – and I think it is linked to the systemic inability of government IT to be agile. Agile is the new black in government IT. However if you are going to be agile then you’ve got to abandon cathedral building

Bureaucracies love building cathedrals, as it creates a false sense of certainty, and of course that suits large systems integrators because they like certainty too. And the thing about building cathedrals is once you’ve started building that’s it.

Except for scope creep of course.

You’ve got the business case, the technologies, the large budget and best of all you can’t stop half way through because then you haven’t got anything. And of course once you’ve got it, you’re stuck with justifying it, maintaining it and the near impossible task of adapting it as needs change

Let me use the Anglican Cathedral in Liverpool as an analogy

Following much debate, church and civic leaders agreed that a new cathedral should be built and in1902 held an open competition to select a design. For architects, this was a very significant event; it to be one of the largest building projects of the 20th century. Two years later they revised the plans to make it more like something else. It wasn’t in use until 1940 and not finished until 1978

To me it sounds like any large government project.

  • Who exactly was the customer?
  • and what was it for?
  • and where was the urgency?

Why do I draw the analogy?

Let’s look at online filing of taxes. Following years of low take up HMRC got legislation making it compulsory. From 1 April this year all companies and organisations must file online and pay their Corporation Tax electronically. However, this week Accounting Web reports that HMRC has delayed April until October. This is because changes in the rates in corporation tax will take that long, six months, to implement in HMRC systems.

Away from the headlines about agility, on the ground, today, it’s going to take that long to change the rate of corporation tax in the system. Some cathedral.

Let’s look at another government cathedral – getting your car tax disc on line. big system. lots of users, seemingly successful

The lead systems integrator is so proud of the result they use it as a case study. but despite it being use to support the case for government agility, it fails one basic test: if they’d been really agile they would have done nothing but thought differently.

If they’d given a stack of tax discs to car insurance companies they could have been online in about a week and they’d have saved the cost of their cathedral except then, of course there wouldn’t have been a cathedral.

And it is the desire to build cathedrals that makes localism so difficult. Cathedrals and centralisation go hand in hand

  • uniformity and conformity
  • control
  • claiming “economies of scale” rather than acknowledging increasing internal transaction costs and loss of flexibility

For a government to embrace localism it will have to stop providing cathedrals and start enabling instead.

Centralised initiatives can too easily displace activty and so destroy or disguise the signals arising in society about what needs to be done and perhaps more importantly what doesn’t need to be done. And that means accepting that different local communities might have different priorities and do different things, and this is where open source software becomes important.

Open source software is released under licences which mean that it is difficult to sell the software at a price above its long run marginal cost of production. Effectively, to coin a phrase well known in the OSS world the software is “free as in beer as well as free as in speech”. This feature can often be dismissed in technocratic terms “yes anyone can look at the source code, but few people will want to”. However, the economic effects are far more significant. The ability to supply the same software and to examine the source code provides “threat of entry” into the market with the concommitant pressure on pricing.

Accordingly, as the market develops the only effective model of IT based on open source software is the one based on local supply as it becomes difficult to hide costs.

A local supplier of services based on OSS is more likely to be able to offer the same for less money than the same from a supplier further away. This market does not exist yet. However the outcomes that might be expected can be identified by analogy. There are several market sectors that exhibit similar characteristics to those that might be expected of a vibrant OSS sector. I decide to choose hairdressing to explain this (it could be building work, car repair or a plethora of services based businesses)

Few people cut their own hair (the equivalent of “examining the source code”) but they have access to the tools and knowledge to do the basics if they so wish. It is a mature market with informed participants. Consumers have access to the information they need to make a decision. Existing and new-entrants have access to an extensive education and training network. (Some of this is public provision, some private or product specific). Knowledge and skills are freely transferable. Success is dependant upon reputation and customer service.

While it is true that customers can choose to pay a premium for a seemingly equivalent experience, it is equally true that they can choose from a range of providers without revealing the basis for their preference. For the most part hairdressing takes place within a local economy. And it it supports and is supported by a local ecosystem.

Local colleges provide skills training, (and require skilled trainers), there are apprenticeships and career progression opportunities. There are a variety of service provision models that can be adopted (salon, contracting, home service, employed, self-employed). Hairdressers require other, ancillary, local goods and services. Value is created and retained in the local economy

As is said of development economics, local people can acquire the organised capacity to define and meet common needs on a sustainable basis.

As the White Paper “Digital Britain” put it

The ability of Digital Britain to contribute its full potential to our future economic growth is critically dependent on having enough people with the right skills in the right place at the right time to develop and apply the new technologies

So to wrap all this up, localism flows from an agility which includes government stopping building IT cathedrals which will require a hard focus focus on open standards and interoperability

Then I can assure you, if open source software is as good as we think it is, it will naturally become the software of choice.

Thank you