Government Digital Service condemned for shallow thinking

I recently described the rightly praised Government Direct website as lipstick on the face of a pig, and gave one example of what lies beneath (there are many more). A reader has just sent me a sharp critique of the Government Digital Services “over-enthusiastic” approach to agile computing. It rather looks as though the reform of public service delivery has become trapped between the Scylla of “delayed big bang” (exemplified by the hard rocks of the original approach to Universal Credit) and the Charybdis of “big bang” itself: sucking departments into a whirlpool of accelerating change using as mix of fashionable approaches from agile to big data, without understanding the disciplines, let alone possessing the in-house skills and experience necessary for their succesfull use on non-trivial change programmes.

In one of my first blogs on When IT Meets Politics I outlined the need to use the classical systems discipline of “structured evolution”, incremental change within an overall framework, in order to achieve the radical changes to public service delivery that are long overdue. Things have moved on since then, particularly with regard to the maturing of Open Source software, but they have not moved on as far and fast as the enthusiasts would have us believe. I have jsut re-read my earlier summary of the some of conflicts that had to be overcome before we would achieve the benefits. We are barely a third of the way along the journey – delayed by unnecessary and unproductive firefights between enthusiasts, like the Open Rights group and those who take a more practical approach to achieving the same objectives.

I am of the opinion that the Government Digital Service (and also the rest of Cabinet Office and Treasury) should focus on mandating inter-operability, at all levels from technical standards and data interfaces through delivery, monitoring and management processes to funding mechanisms, rather than particular approaches or methodologies. Modular inter-operality is central to enabling flexiblity as needs and organisational structures changes and competition between suppliers. It removes the need to plan ahead in impr5atical detail and also enables the removal of contractual lock-ins (“components” can be replaced) which is why so many suppliers who “talk the talk” are very reluctant to “walk the walk”.

It does, however, require mandation. I was supposed to use one of the early “agile” technologies employing re-usable code modules in order to merge and decimalise the sales ledgers for the ICL group of companies. I cheated and wrote custom code (albeit in well documented “common standard” Cobol) to cut days off the conversion and hours off the weekly overnight run for the new system. I was, of course, found out during my first test runs: the operators reported me even before I was ready to confess (it was a well run department). But no-one disciplined me or ordered me to follow process. Instead I put a nail in the coffin of that particular set of “agile” tools. Supposedly the technologies have moved on over the past forty years – but have the mindsets of young enthusiasts – as I was then.          

More-over I benefited from training and career development programmes, including structured and supervised, “apprenticeship-like” work experience and a subsequent full-time MSc at London Business School. Few, if any, of the public servants of today have had anything like those advantages. Hence my regular plea to implement the recommendations of the Fulton Report . Relying on private sector expertise is no substitute and the current Civil Service Learning framework is a joke in very bad taste.  

P.S. Just been sent a link to the interview with Brian Wernham on BBC News 24 on the (finally) phased and incremental implementation of the Universal Credit. Do watch. As readers will know I have been blogging on the need to adopt such an approach for over two years and have just looked at  again at one of my first postings. Brian was diplomatic and did not comment on the aim of completing the transition by 2017. I would merely say that having the new systems ready by April of next year for a straight line roll out at 300,000 a month is improbable as a successful way forward. Much will also depend on the success or otherwise of the switch to RTI for PAYE. We should remember that  PAYE is itself a creation of wartime, when the entire nation was mobilised working for the war effort except for those running the black market. The re-imposition of that mindset when a growing number of us no longer have one job at a time, let alone for life, is “at least as ambitious as the Universal Credit”.