Going outside low-code silos for Mutable Action Automation - Your toolbox needs low or no-code solutions, even if they are called RPA or Rules – or Uniface
Published:
Content Copyright © 2022 Bloor. All Rights Reserved.
Also posted on: Bloor blogs
I attended a Uniface Virtual User Group Webinar a few days ago and I was impressed. It has been acquired by Rocket, but many people I recognise are still there and the roadmap under Rocket looks good, with increasing support for microservices and Docker. I’ve always had a soft spot for Uniface since I saw an attempt by a business group to use it back in the last century undermined by an IT group. Sometime later I had the chance to ask the IT manager responsible what that was all about and was told that the potential deployment had to be sabotaged, because the Uniface low code solution fitted the business needs so well, and cost-effectively, that it would make the IT group’s official C++ “object oriented” development strategy look silly! Yes, Uniface is that “old” – perhaps “mature” reads better but it has always taken the approach of letting customers upgrade at their own rate, and not forcing upgrades on them; and its separation of business logic from the user interface has allowed it to modernise its user experience without significantly impacting existing Uniface solutions. It has a long history of being acquired by various players (it has a healthy and happy customer base, typified by companies like Pernod Ricard, which gives it value) and I took an interest in it again back when it became an independent Netherlands company for a while.
It was acquired by Rocket Software, back in 2021, and it should be a good cultural match with Rocket: both companies are long-established and focus on customer experience. This comment by P. Gary Gregory, President of the Database and Connectivity Business Unit at Rocket Software, fits well with what I know of Uniface too: “This isn’t just a technological match, but also a philosophical one. Both companies are in their fourth decades and are recognized leaders in their respective fields who share a passion for putting their customers first. Rocket extends our core values of Empathy, Humanity, Trust and Love to Uniface customers and partners. We are excited to welcome a great new team to Rocket as we continue to expand our business around the world by creating legendary outcomes”.
Model-driven, low code platforms are not new – and there are increasingly many examples to choose from – which means that products like Uniface are tried and tested in the field and are proven to scale well. Actually, it’s not just scalability that matters, but achieving effectively “no limits” – not just limitations in the size of applications, but no limitations around business complexity either, as if there are limits you will hit against them. Uniface also supports the “start small and grow with success” implementation model – it has a free community edition – which is very welcome. It is rather interesting that one of its current competitors, Mendix, also started in the Netherlands.
Do I have any concerns about considering something as established as Uniface? Well, not around its age, Uniface hasn’t stood still and has actively modernised over the years. I suppose that I prefer “no code” to “low code”, and Uniface is definitely “low code”, but its productivity gains over conventional procedural programming are easily sufficient to keep it in the game (and some people prefer low code to no code anyway). I suppose that standardisation and interoperability are the elephants in the room for non-procedural higher-level languages – how proprietary are they and how easy is it to define and test an exit strategy – but, although an issue to be aware of, I don’t see this as an especial issue for Uniface. Especially as it has put a lot of effort into making its developer user experience reasonably familiar to Microsoft/Eclipse developers.
What I do see as an issue is the possible siloisation of development approaches. Even conventional coding has a place (access to libraries of pre-built components makes it more productive these days), even if I think that low- or no-code approaches are usually more useful, largely because they let developers focus on solving business, instead of coding, problems (although their productivity plays a part here, too). But what of rule-based approaches? Robotic Process Automation (RPA)? Workflow? ERP (which may include RPA and workflow features)? And so on. These approaches are often treated as incompatible silos and if an organisation adopts several of them, using each where it is most appropriate, the teams involved probably don’t talk to each other (especially if a business group acquires a no-code product without talking to the IT group). This can result, at best, in the use of less effective development technologies; at worst, to an organisation that isn’t agile and can’t manage effectively the changes needed to accommodate its business customers. As an aside, this doesn’t mean that I’m against end-user IT, “grey IT” or whatever – just as long as the IT group is aware of its use and has appropriate input. And doesn’t sabotage non-code approaches with a “not invented here” attitude.
Bloor believes in the Mutable Business, in a constant state of evolution. Evolution means that applications (or, at least, their associated user experiences) are constantly changing – and the business can’t afford to waste time on changes that aren’t directly useful for the business, don’t work, or which take several iterations in production before they deliver anything useful to the business. A change in the business, or an insight from analysis, results in a proposed Action which must be Automated (manual implementations, in general, are error-prone and don’t scale. So we talk about Action Automation, and about using the most effective tool for the automation. If development is silo’d, there is a strong risk that the most appropriate tool won’t be used.
That said, since many Mutable changes will be around user experience and need to be implemented quickly and reliably, low code approaches like Uniface (and others, such as Mendix and OutSystems) might be among my first ports of call. Just not necessarily my only ports of call.