Saturday, 9 July 2016

To code or not to code

There's an ongoing debate on the LocalGov Digital Slack Team about the merits of a low code approach to building digital services.

A low code platform is one that enables you to quickly create and deploy digital services usually by dragging and dropping a selection of predefined plug and play elements. These could be anything from a text input box, to a database or API call, to a bit of logic based on answers to previous questions.

Where I work we've been using a low code platform for 3-4 years and some of the stuff you can see being used in our dashboard was built without writing a line of code.

There are great benefits to having this approach available to you. Just one example of this is we're currently user testing an end-to-end register for the Self-build and Custom Housebuilding Regulations which we built in a couple of weeks. We'll make this available for anyone using the same platform as us, for free.

Because we're starting to use the Local Government Digital Service Standard and therefore the Government Digital Service design patterns, when we needed to include something that checked whether someone was over 18, we built something that looks like this. The great thing about a low code platform is now that we don't need to build that again. We can drag and drop it into any new digital service and it'll work straight away.

So low code is the way forward, right?

Well yes and no. If you're astute you'll have probably guessed that when we built our over 18 checker, we did that by writing code. The design pattern suggests that individual inputs for Day, Month and Year are best for usability for memorable dates, and we didn't have anything that merged these together into one date, then worked out if that date was more than 18 years in the past.

When you start to think about it, actually every digital service runs on code and markup, even if you don't write it or see it yourself. So actually low code is really low code once you've written the code.

Through having access to this method of building services over the past few years it's become apparent that as digital services evolve, and when you start to tackle some of the bigger transformation work, there's an increasing need to write connectors, end points and other more bespoke functionality. This isn't a one off process because the services your organisation consume will change, or you'll want to change how you use them to respond to user needs,

Without a developer you'll be at the mercy of a supplier, and you'll be in the same position some have found themselves in the past, with a supplier wanting to charge £100s for a minor change to the HTML of a web portal. This isn't prudent use of public money, nor will it enable you to get something quickly most of the time.

So my advice? Low code is a great idea and will help you accelerate the creation of your digital services. Do seriously look at it, but don't outsource your ability to create and adapt your re-usable low code elements yourself, or you'll find yourself a less agile and adaptable organisation.
This blog is written by Phil Rumens, Vice-Chair of LocalGov Digital, lead for LGMakers and who manages the digital services team at a local authority in England.

The opinions expressed in this weblog are my own personal views and in no way represent any organisation I may have worked for, currently work for, might be thinking of working for, might not be thinking of working for or have never worked for. In fact having said that they, might not even be my views any more as I might have changed my mind so I wouldn't take any of it too seriously.