Serverless’ greatest strength is also its greatest weakness
We’re big fans of keeping track of what is going on in the developer community. Today, we’re taking a look at SlashData’s most recent survey of the developer landscape. So, what does the technical world look like today? And more importantly, where is it going?
SlashData’s Developer Economics global survey reached more than 21,000 developers from around the world and focused on four major themes: AI, serverless, augmented and virtual reality, and programming languages.
According to their research, Machine learning and AI are poised to fuel a new wave of innovation. This shift requires new and more diverse skillsets from developers, including a background in engineering, biology, mathematics, probability, and statistics.
Spurred onwards by advances from neural networks, improved processing resources, and easy access to data, machine learning has developed considerably in the last year. Over a third of developers surveyed believe that machine learning is the next big thing.
However, there are challenges that lie in wait. One of the biggest problems facing machine learning models today is the inability to retain, pass on, or apply knowledge gained in one process to the next. Context and relationships are lost. As a result, the machine learning models and architectures that are emerging require information retention and improved ability to transfer this knowledge from one process to the next. This is still in the early stages, so only time will tell.
Adopting of serverless platforms continues to grow quickly. 19% of backend developers are now using serverless platforms, up from 16% half a year ago. This is nearly equal to the use of VMs (20%) among backend developers.
AWS Lambda is the undisputed champion of serverless computing platforms – 44% of developers use it. 25% of developers use Microsoft Azure and 24% use Google Cloud for a close third. There are a number of smaller fish in the serverless pond (Twilio, IBM Bluemix, and Alibaba Cloud, to name a few), but none of them have managed to grab more than a 6% market share.
The study points out that serverless’ greatest strength is also its greatest weakness. It can be adopted in a piecemeal fashion, solving individual problems. However, deeper adoption is likely to lag behind, as the serverless paradigm forces a microservices architecture. Many developers won’t be able or willing to migrate wholesale to one platform.
Figuring out how popular a programming language is hard. There’s an awful lot of metrics out there and we keep an eye on all of them, including TIOBE, PYPL, Redmonk, Stack Overflow’s annual survey, and GitHub’s yearly analysis of the Octoverse. In this case, the State of the Developer Nation went over active software development via an independent estimate of software developers and a biannual survey.
Java has retained its popularity at number two, likely thanks to its lively and healthy mobile development ecosystem with Android. However, it is considerably less popular for tech like IoT. Additionally, Python is rising up the ranks, most likely due to the rise of machine learning. Niche languages like Swift, Ruby, and Lua seem to be holding steady without much growth.
SlashData’s research has confirmed a particularly thorny issue in tech: older programming languages tend to garner vocal criticism while new programming languages have enthusiastic support. This emphasis on novelty makes it hard for new languages to really grow too much, before they are discarded in favor of something shiny and new.
Conclusions for developers
While this research may not have documented any revolutionary new trends about the tech industry, it’s always good to get confirmation that we’re all going in the right direction. From advances in artificial intelligence and machine learning to serverless capabilities on AWS, the top trends in the developer community are clearly defined.