By Rod Cope, CTO, Perforce Software

The Future of Software Development - A Look Ahead to 2025

There are lots of great commentaries covering software predictions for 2020. This post looks at 2020 and beyond to focus a longer lens looking at what is going to happen in software's evolution between now and 2025, and the impact that will have on software development.

1.      Scale takes on a whole new meaning: Analysts are saying that we are going to see 10 times more change than we have just gone through.  The number of Internet users worldwide will double, cloud and ‘everything as a service' will become even more ubiquitous. 

2.      We are truly in an app-driven world: It took 40 years to create the first 500 million apps, but it will only take four years to create the next 500 million.  However, those apps are going to be a lot smaller, faster, more agile and involving more moving parts, as well as being event-based.  Apps will depend on hundreds, even thousands, of APIs.

3.      Hardware disappears into the background: Hardware is going to become the wrapping around software. Take for example autonomous vehicles, which are essentially computers inside a chassis.  There are already over 25 companies trying to create self-flying vehicles now; imagine the software that involves.

4.      Quantum computing is a game-changer: Rightly being called revolutionary by Gartner, quantum computing is going to enable so much innovation that is possible today, and help overcome some of the complexity and volume challenges that exist today.  Volkswagen is already working on a city-wide traffic management system based on quantum computing to orchestrate every car, pedestrian, light, and other elements.  There is already talk about quantum medicine, with DNA-specific drugs.  Who knows what other possibilities quantum computing will bring?

5.      Personalised and predictive: Software is going to become less reactive and more predictive, largely driven by AI.  Perhaps the best way to explain this is through an example.  A keen hiker asks his smartphone for suggestions for new walking boots.  The smartphone suggests a pair, based on the context of knowing what the hiker spends on shoes, what style, what his friends on social media like, what he has historically looked at or liked.  The hiker says ‘OK, show me those in brown,' so the smartphone renders an image, the hiker says ‘Yeah, I want those.' The smartphone goes ahead and orders a pair, together with some socks, and put it on the hiker's Apple account so that he gets some iStuff points.  Everyday tasks like this example - whether at work or outside - will just happen.

What This Means for Software Developers

Software may be eating the world but, to paraphrase a quote from Nvidia's CEO a couple of years ago, AI is going to eat software.  The implications for the future software industry workforce are massive. While AI itself might lead to job losses, there will be a huge demand for anyone with AI.  Grab anyone with AI specialisms now, even if there is not yet a clear role for them: there will be very soon. 

Existing developers will need to retrain and there will be a lot for them to take on board.  They will need to have greater breadth of knowledge but, to stand out from the crowd, it is a good idea to still specialize in one area: a jack-of-all-trades and master-of-one scenario.  New tooling, based on AI, open source, cloud, and containers will help make job development easier.  For instance, junior developers can now carry out tasks previously restricted to senior developers, helping a little to relieve some of the pressure of the developer shortfall.

Humanity 2.0

Plus, humans have a lot to offer: people are very good at taking in data - about 10Mbps mainly through our eyes and ears - and will soon use tools that let us interact with the world just through the power of thinking.  Developers and others who use these new capabilities to improve user experience will be in high demand for many years to come.

DevOps at Scale

The past couple of years have brought a realization that the volume and complexity of software means finding better ways to deal with scale, while maintaining quality and security, and without sacrificing speed.  With the proliferation of APIs, finding better ways to manage and secure them across the API lifecycle becomes essential.  Poor API management is already one of the reasons that many organizations are finding microservices hard to deploy effectively, but the tools are there to solve that challenge.

Automated and continuous testing is already common, and we now have smart testing using ML/AI technologies, which further reduce brittle tests that break every time the UI changes.  Alongside testing sit automated tools that continually inspect code for quality, compliance, and security.  Without these foundations, the risks around software that moves so fast, has so many elements, and is connected to so many third-party sources, become alarmingly vast. 

This is a decade that brings huge technology opportunities, together with some big challenges, which is why it is important to start planning ahead and take control now. To borrow a quote from the Terminator films: "The future has not been written. There is no fate but what we make for ourselves."

##

About the Author

Rod Cope, CTO, Perforce Software

With 25+ years industry experience, Rod Cope is CTO within Perforce Software, and provides technical vision and architectural leadership for the company's globally distributed teams. Rod was the Founder and CTO of OpenLogic and joined Rogue Wave as CTO following the acquisition.  For the past two decades, Rod has spoken on a wide variety of topics at events around the world. Apart from API Management and security, other topics he has commented on recently include micro-services and multigrain services, Agile methodologies, open banking standards, digital transformation, and software development trends in general, especially at scale.