Type “the future of programming languages” into Google and it generates approximately 180,000,000 search returns.
You can read countless blogs debating the pros and cons of popular programming languages and which ones will dominate IT over the next decade.
For IT directors and chief operating officers, much of this discussion about the perfect programming language and pristine code is of limited interest. The thing that matters is whether the programming code is reliable, fast and flexible. Are IT projects finished on time and within budget? Does software performance meet service level agreements?
The future of programming isn’t about searching for the pristine and perfect code, or about a new programming language. It’s about writing less code − or in some cases, no code.
The most important trend in programming for the next decade will be using machine learning and artificial intelligence to automate much of coding.
AI and machine-based learning can automate coding and help programmers write faster and better code. When combined with intelligent “Integrated Development Environment” — software used by programmers when they are writing code in a specific platform, such as Visual Studio etc— it can save companies a lot of time and money.
In June, Facebook researchers said that they had developed a new type of technology to convert code from one programming language into another. The technology, known as a “transcompiler”, converts code from one programming language like DELPHI or PASCAL into C# or JAVA.
When converting code, it’s easy to miss business rules − for example, for a bank, stipulating that a bank account can only be opened after certain ID and other compliance checks had been done. That can really screw up the new code.
It can also be expensive. As Facebook researchers noted in their research paper, the Commonwealth Bank of Australia spent approximately $750 million and five years converting its platform from COBOL to Java.
So, IDE’s can have in-built feature of transcompliers. Suppose a programmer selects a block or code or a class and use integrated transcomplier in the IDE to convert the code. This will save a lot of effort, time, money and no business rule will be missed. (Fig 1)
The savings of time and money will depend on the application and complexity of its code, but organisations could certainly save weeks or months of time — and millions of pounds for large IT projects − by using AI-based programming and code conversion.
5G and programming
The other big change in programming over the next decade will be 5G. Increased network and computing capacity where you can download a HD movie in just 2 seconds. It will lead to changes in programming languages and new programming languages that can capitalise on the computing power of 5G network and build software applications using the network, including for transformational projects.
5G will be fast and powerful enough to take augmented reality, virtual reality and mixed reality technologies to the mass market.
In the business to business market, this could mean a factory technician can get a support through a VR/AR headset without any delay or buffering instead of providing support over a phone call.
Similarly in healthcare sector, it could mean a doctor using a video streaming and a robotic arm, combined with augmented reality glass, to examine patients remotely. This could improve the healthcare scenario of developing nations where they still lack medics in rural areas.
Programmers will still need to learn new languages over the next decade. However, their priority and their employers’ priority should be getting a good grasp of the main IDE systems, machine learning and AI, and how 5G will drive new programming language and software applications.
We are moving to a new IT era from high code to low code or no code. AI and automation of code conversion can speed up programming and IT projects. If the end code is good enough for the software application, who cares what language it’s in.