Recent history of UK government and open source

Over Christmas I spoke with a team in the US government who are pulling together some work on open source policy on that side of the Atlantic. To help them I tried to document recent UK government history on the topic. Having done that it seemed helpful to publish it somewhere in case I ever need to referene it, but the GDS blogs didn’t feel quite right.

This is definitely incomplete and I know a lot of other people were doing a lot of work. There’s a clear GDS-centric slant here because that’s what I know first hand. If you spot any particular egregious missing pieces, feel free to use the comments to add them.

In the first couple of years of our last Parliament (which ran from 2010-2015) there were two parallel strands of work running which eventually came together in GDS. The first was focussed on ICT/”technology”, and the latter on “digital”.

The last government’s ICT strategy laid out the initial commitment to a level playing field between open source and proprietary software

Most of the ICT strategy’s recommendations are embedded in the “Technology Code of Practice” which we still use, and which is backed up by a set of spending controls.

We started our work on GOV.UK in the open but didn’t have the support in place to really support any of that code by packaging it for others’ use. I coined the phrase “coding in the open” to describe what we were doing and blogged about that.

The Design Principles we published early in the life of GDS have been the rallying cry for openness in everything we do.

The Open Standards Principles were one of the areas where the digital and technology groups began to come together (we formally merged into GDS in 2012).

The Digital by Default Service Standard was gradually piloted over 2013 but came into force in 2014. It’s been revised a little, but point 8 has been consistent.

Our “Service Design Manual” includes some content on open source – that’s still weaker than I’d like in terms of advice and we’re working on an updated version that should begin to appear during 2016.

The project GDS has invested most in open sourcing (to date) is called vcloud-tools. My colleague Anna wrote about our process for that.

A lot of this is underpinned by procurement reform that intends to make it easier to buy smaller pieces, work to open standards, etc. Examples of that are our G-Cloud and (forthcoming) Digital Outcomes and Specialists frameworks:

I see a lot of what’s happened to date as being about laying the groundwork for what we really want to do, which is about building a community and leveraging this work to reduce duplication, better understand opportunities for consolidation (eg. providing more common platforms), and also softer things like improving our profile as an employer of technologists.

We’ll be gearing up to really invest in that in the new year, but we’re doing a similar thing around the “open standards” side and some recent blog posts start to give a flavour of that.

2 comments

  1. James, just because I was mailing you anyway, I thought I would drop a comment here.

    What GitHub sees is the following:-

    1. Significant use of GitHub to publish open source components to government projects. This is great, but (so far) has seen a limited volume of opensource community/citizen contributions – civic hackers are just not engaging on UK Public sector projects.

    2. OpenSource contribution underpinned by UK Public Sector investment. Unsurprisingly, this is close to non-existent. The reasons for which are well documented. Here are some notable exceptions:-
    Gaffer – In December 2015, GCHQ OpenSourced one of it’s in-house tools, a Graph Database called Gaffer. Moreover, it seems to be actively participating in its maintenance
    Code4Health – NHS England’s program for community software development for the healthcare secot
    * IMS MAXIMS – Electronic Healthcare Record funded by the NHS (Taunton & Somerset NHS Trust)
    * RippleOSI – Funded by the NHS Tech Fund, and hosted by Leeds Council, this is an open source patient healthcare care record optimised for social care.

    What we see with the active program taken by NHS England to promote OpenSource are the tensions inherent in public sector OpenSource adoption:-
    * Firstly the vast majority of funding is still going to proprietary technologies (as illustrated by the Integrated Digital Care Technology Fund – where spend is still mostly going into projects like Cabridge’s eHospital program).
    * Secondly, where acceptable/adotable OpenSource projects do exist, adoption is being blocked in the name of supportability
    e.g. RippleOSI (with overall NHS development funding of £4mil) was considered for a recent contract award by another council, but could not meet the challenges of guaranteeing supportability into the future – the contract was awarded to another SII (at a cost of £7.4mil over 5 years).

    3. UK Public Sector ability to
    (i) Consume OpenSource
    (ii) Develop code in the same way that OpenSource projects work

    Digital skills are merely the start of the matter. Inserting developers and digitally capable personnel into the organisations do not necessarily change outcomes. Firstly, Conway’s law comes to the fore – organisations engineering efforts bear a resemblance to their own structure (see this article by the team at GCHQ (bless them for publishing on GitHub!).

    We would argue that lessons need to be learned by the actual dynamics of OpenSource development. That organisations who can successfully collaborate in the bazaar type methodology of opensource (see research by Klaas-Jan Stol,) can produce code more efficiently and dynamically. This process, clustering around the term “InnerSource”, is leading to an evolution of software engineering now going prime time in places like PayPal and Bloomberg. My (personal) conjecture is that UK government looked at OpenSource as a method of (license) cost reduction, but have a long way to go to realise the full benefits that the OpenSource community can teach it about software engineering (in my experience, benefits realised wholeheartedly by the (limited) DevOps community in government (and some CDO’s), but still a long way from the agenda of most IT, and nearly all executive, leadership within the civil service.

    That sounds like a “Top Gear” bombshell moment, and one I would be delighted to discuss. I hope this gives more material for thought.

    • Thanks Alex – lots of interesting points there.

      It’s probably worth being clear that my experience is very much around UK central government so I don’t have all that much insight at the moment into what has happened in the health sector.

      There’s clearly a lot more to be done, though it’s worth noting that:

      * There are more contributions than it’s easy to identify, partly because developers are quietly getting on with submitting patches back to a range of projects. In many ways it’s a good thing if that can happen in an unsung way as it means that people are empowered to just do the right thing. There are also other projects like vcloud-tools that have had external adoption.

      * Comparing investment is a very difficult task given the very different funding models for open source vs. proprietary software.

      We’re in the process of appointing a new leader to coordinate and amplify all our work around open source. I’m hoping someone will be in post in the near future and we’ll be able to pick up the pace on many of the areas you mentioned. You can see the job description I used for that role at https://github.com/todogroup/job-descriptions/blob/master/uk-gov-gds-open-source-lead.md

Leave a Reply

Your email address will not be published. Required fields are marked *