I’m a member of the Cabinet Office Skunkworks project ‘dotgovlabs‘, where ‘innovators’ post ‘solutions’ in different ‘challenge’ areas. Rather than write anything more (that you won’t read) about dotgotlabs, here is a video produced to explain the project … including me and others with great faces for radio.

With interests in open source, open standards and interoperability and with experience in and around the NHS, I’ve contributed to many solutions posted in the Healthcare challenge.  I’ve commented on the interoperability challenges facing ‘Solutions’ such as ‘Give Patients Access to Data’ (a challenge involving Patient Opinion and MyDex, see previous post on Brent Council opening up resident data access) and on the use of social media in healthcare solutions … and the privacy implications there in….

I’ve been contributing heavily to a solution posted by Tim Knight which proposed greater use of open source practices in the NHS – provisionally titled Open Me Project (login req; email me for invitation).

The goal of the solution is to provide a framework and supporting tools for the collaborative development of open source solutions for NHS organisations and commercial healthcare providers. The solution has been voted up to progress from ‘Buzz’ stage to ‘Teaming’ stage.

As part of the incubation process for solutions, DotGovLabs have engaged the support of Experts from Departments of state and the Private Sector to give feedback and pose challenging questions to teams. This weekend David Ostler, Chairman of unitedHealth posted feedback on our solution and several others.  I thought I would repost the exchange here.

David Ostler: “What are the reasons why open source adoption has lagged. Been around for a long time. Why hasn’t it taken off – seems like it should be more widespread than it currently is. Must be reasons why? What are they?”

Rob Dyke: “Hi David, thank you for your questions. I’ll make some answers specifically around NHS/Public Sector IT, before looking at open source more widely.

I think that the Cabinet Office identified the reasons for the lagging of open source software in its recent ICT strategy document: the playing field is not level. Two significant factors skew the market.

First, the trend towards ever greater outsourcing has diminished the ‘intelligent customer’ capacity in the public sector, leading to procurement by tender using risk based assessment models. This inherently favours proprietary software from corporations with the resources to hedge the risk. It matters not whether iSoft deliver what they are contracted to do if there are commercial mechanisms of compensation and insurance to protect all parties. When the relationship between a no longer ‘intelligent’ Customer and BIG IT Service Provider is asymmetrical, a DoH/iSoft outcome is guaranteed.

The second factor is one of scale. There was for some time – and still is, in many respects – a desire for BIG IT. Capital I, Capital T projects would deliver all singing, all dancing, pervasive, unified processing environments for all transactions the Secretary of State for Department of X could dream of. This again favours proprietary software from corporations with the resources to speculate on a tendering process where the financial rewards are great. Although why proprietary software should be favoured by public sector procurement is a mystery to me. It would make more sense for the public sector to be an equal owner of any intellectual property created, at least until delivery (see this Register article from 2005; six years on ‘moral hazard’ is still a crucial question for the DoH CfH NPfIT relationship with CSC, iSoft and BT.).

These two themes were consistently and repeatedly given in evidence to the Parliamentary Public Administration Select Committee on ‘Good Governance: The Effective use of IT‘ and were central to the argument in the Institute for Government‘s recent report ‘Fixing the flaws in government IT‘.

The Cabinet Office have set out the operating framework for public sector IT in the Strategy published this last week: more Open Source and no more BIG IT. The strategy is welcome, however I will wait for some of the goals to be achieved before loudly heralding in a new era.

Looking more widely than public sector & NHS IT, the adoption of open source and open standards has been the crucial factor in the growth of internet. From the protocols driving the networking layer, to the standards enabling the presentation layer, open standards, especially those embodied in open source software have scaled to the immense size of the internet. Open source software is used on more webservers than closed source alternatives, as the NetCraft ‘what’s that site running‘ survey shows us month in, month out.

There are many talented development teams in local information and infrastructure teams across the NHS. Supporting this great pool of talent with tools like the collaboration / development forge proposed in this Solution will support the Govt ICT Strategy from the Cabinet Office.”

David Ostler: “Completely agree with the comment about open standards. These are required in a public utility like health care. (my opinion is that the NHS does this better than most health systems).

I still can’t figure out why open source is used less than you think it ought to – I am not an expert there. But in the for profit world I live in, if I can use open source to drive down my cost of development, I have complete incentive to use it, It is lower cost – and I either increase my profit or have a lower price which means I win the bid. We do use it, but not close enough to the decision making to know when and why we chose to use it. I am sure we consider availabity of skill sets, scalability, supportability and total cost of ownership.”

Rob Dyke: “On open standards. I agree: The NHS excels in this area. You only need look at the CfH Data And Standards Team for some great examples of definition and ownership of standards.

On ownership, value and accounting. I was deliberately careful not to refer to ‘free software’ nor to get into TCO discussions. This is because the truth is open source software has a cost, freely downloadable or otherwise, just like any other software. Software needs implementing with a framework of infrastructure and support with relationships to the vendor or developer in order to resolve issues that arise in the environment. This all contributes to a TCO of an technology environment; the capital expenditure of a license for commercial proprietary software is just one part of the cost calculation.

The phenomena of the commercial open source service provider is relatively new in comparison to the longevity of the Microsoft’s and IBM’s of this world. Commercial open source service providers have a different production logic to the traditional BIG IT vendors. Consider Redhat, the $1bn revenue / $8bn valued software services company. Redhat sell value add services and support around an open source operating system.

Or Alfresco, who develop and maintain an open source document management and collaboration suite.  Like many open source vendors, Alfresco offers customers a Support Subscription which provides users access to:

  • Technical Support Services
  • Maintenance fixes, patches, and updates
  • Engineering support / Escalations
  • Alfresco Network – End user support portal
  • Certified stress tested builds
  • Access to the Alfresco Enterprise software

My point being two fold: the first being that open source businesses operate outside the Fordist production line model of software production as a captial project. Rather the product is incrementally developed and expanded, usually with a ‘crowd sourced’ dymanic. Commercial open source businesses supply support services and priority access to engineering which are chargable, while the software is free (at least, the community editions are free and the Enterprise features available in a subscription). The second being that there is a ‘for profit’ dynamic in open source software production, just as there is in proprietary software production.  You just need to look at what is being sold. Spend a few hundred quid with Microsoft and see what support you don’t get from no access to an engineering team. Then spend the same few hundred quid with Alfresco and see what a difference support makes.