OutSystems 9 Amsterdam – a consolidation release

Written By:
Content Copyright © 2015 Bloor. All Rights Reserved.
Also posted on: The Norfolk Punt

On the face of it, OutSystems 9 Amsterdam is a consolidation release for Outsystems Platform – yes, there’s some nice stuff about scaling and integration with other platforms, but I thought it already did that pretty well (although improvement is always welcome, of course). It will continue with the picturesque release names but (unlike with some others), there is order in the madness: after Amsterdam, the next point release will start with B, then C and so on. I’m in favour. See here for release details – presented as structured web pages instead of as a massive PDF (or, worse, txt) document. Again, I’m in favour.

Nevertheless, if you dig a little deeper, there is some seriously interesting stuff in this release. For a start, OutSystems seems to be taking mobile offline really seriously (its own research suggests that mobility is seen as the business functionality most critical for apps in 2015). This is technically challenging, because if two people make clashing changes while offline, resolution of the clashes can never be totally automated; although technology (done well) can minimise the impact of this. However, worse, it involves people issues and politics – in the case of a clash, who is allowed to prevail? Outsystems is providing templates for “good practice” offline working, for use cases with different online/offline priorities, in its Forge community (where customers can contribute too). Living in Wlltshire (where mobile broadband is erratic, to say the least) I regard mobile offline capabilities as very important to any mobile capability; and think that providing good practice templates for its implementation as an extremely effective approach to ensuring that it is done well.

Another area that I think is increasingly important, is the user experience. Here, in the Amsterdam release, OutSystems has extended its visual approach to replace Javascript, HTML5 and, especially, CSS programming with something more productive: the Silk UI open-source framework, which uses themes, patterns and samples to make it easier to, as Outsystems puts it, “build complex user interfaces… which look beautiful on any device”. This is part of OutSystem’s new Ominichannel Experience and another feature of this I like is a standardised way of collecting end-user experience feedback – which should help to ensure that you are rapidly delivering “user delight” instead of just more applications that don’t quite do what the business wants. Or, at least, it should ensure that if you do go off-track, someone tells you promptly – its a governance thing.

As for scalability, yes, OutSystems Platform does play with the big boys (and has done for some time) – with Hadoop applications as well as with mainframe database applications. Even so, since scalability is one area where other RAD applications have fallen down in the past, improvements are always welcome; and as many large-scale use cases as it can find will help OutSystems no end. And, in fact, Sean Allen (OutSystems’ Director, Product Strategy) did find me several case studies that I’d see as confirming OutSystems’ ability to scale. Liberty Insurance, for example, with about 12 developers and 39 deployed applications, delivered over about 10 years. OutSystems says that “Liberty’s IT team owns and maintains one of the most complex OutSystems ecosystems in the world. Liberty Connect has become business critical. It interconnects all systems of record allowing everyone to collaborate through a single, tailored solution” – and includes links to DB2, Oracle and SAP. Then again, the OK! teleseguros e-commerce insurance site claims high scalability, with 160,000 customers in total and 800,000 unique visits in the first two months. There are others on the Outsystems site, here; if you do have a very large scale application (and “scale” isn’t just data volumes; it’s, for example, the number of reliably near-real-time responses per hour and the number of concurrent users, too), you should probably spend quite some time reviewing the approaches taken by other people to similar applications; and make sure that you design for scalability from the first.