2016 in Software
Microservices, Culture, Papers, Data Science, Open-Source .NET, VR/AR href> 2016 was the year that the Microservice architectural style went fully mainstream. Its attractiveness to both developers and DevOps has made this the de-facto standard for new development in many organisations. As part of that transition, we’re seeing a gradual maturity in the framework choices available. As Adrian Cockcroft pointed out in our podcast back in April, the actual choice you make is heavily dependent on the languages you use. For example Netflix, the original “poster-child” for the movement, is Java focused and, through their partnership with Pivotal, have helped to ensure that Spring Cloud and/or Cloud Foundry represents a good choice for Java shops. Broadly, it feels like we’ve settled on a fairly standard approach for building most modern software; chances are you’ll be using one of the agile methodologies to manage the product, using a container such as Docker to package it with a PaaS as the deployment target, DevOps to run it, and Microservices as the common architectural style. An interesting aspect of this is that as this combination moves from start-up to the enterprise, we’re needing to find ways to scale from being a team-based optimization to an organization-wide discipline. Related, we are seeing Bay Area companies popularising the idea of the “you build it, you run it” style of team autonomy, and at a higher level, intentional company culture is becoming a key differentiator for start-ups looking to hire and retain talent. We launched our culture podcast specifically with that movement in mind. Perhaps driven by the idea that every system is distributed, 2016 also saw a growing interest in collaboration between industry and academia. Kolton Andrus and Peter Alvaro spoke about their collaboration around failure testing in a keynote at QCon London, which was one of my favourite talks of the year. A related trend is around reading and discussing academic papers, with the “Papers We Love” chapters leading the charge. We were thrilled to host a chapter at QCon New York, and in partnership with Adrian Coyler, have launched our own regular applied computer science eMag “The Morning Paper.” A third major trend in 2016 was all about data science. We saw tremeondous interest in streaming, with our ongoing series of articles on Apache Spark doing particuarly well. 2016 was also the year that machine learning and Artificial Intelligence started to receive mainstream attention. An exciting part of the trend is that many companies are releasing sophisticated libraries and tools as open-source, making them available to a wide variety of developers. Google’s Tensor Flow has captured much of the early interest, and Microsoft’s Vowpal Wabbit is also getting attention. John Langford told us in August that the news recommendation system built on Vowpal Wabbit has been deployed on MSN, and led to a 25% improvement in reader engagement. At InfoQ we are also experimenting with machine learning for content recommendation and are encouraged by the early results we are seeing. If you would be interested in joining our closed beta program in early 2017, please reply to this email and let us know. For all that data science algorithms can be effectively applied to a variety of machine learning problems, the misuse or misunderstanding of the technology comes with wide-reaching and alarming implications. Cathy O’Neil’s book “Weapons of Math Destruction” (review, podcast), is an excellent and accessible introduction to the dark side of big data. With increasing concern around personal privacy, firms such as Apple and Google are turning to differential privacy, a cryptographic technique which introduces a low amount of noise into the data to prevent identification whilst still retaining statistical accuracy. Apple, in particular, is to be applauded for its stance, but it should be noted that this is still an ongoing area of research. If you are new to the field of data science and want to know where to start, our article series could be a good launch pad. We’re planning a second series on machine learning for early 2017. The .NET space saw somewhat of a resurgence, as a result of its push into open source. From Microsoft, we see the new cross platform efforts coming together in a new family of products, reaching 1.0 (ASP.NET Core, EF Core, and .NET Core). Outside of Microsoft, we'll see developers leveraging .NET in big and small ways. We’ve run .NET tracks at all three English QCons in the last 18 months and have tracked progress closely on InfoQ. We recently ran an eMag on What to Expect in C#7 which did particularly well for us. It was a busy year for JavaScript, with significant interest amongst InfoQ readers in Angular 2, Ember.js, and React.js. If you are new to front end development, Bonnie Eisenman's "Map for Newcomers" is recommended reading. Finally, we see augmented and virtual reality generating business interest with hardware such as the Oculus Rift, HTC Vine and Microsoft HoloLens, that are maturing to the point where early adopters can start to explore real potential applications beyond games and novelty. InfoQ sees applications for VR for more empathetic remote collaboration, but the technology feels a way of being ready for mainstream adoption. Amber Case spoke about VR at QCon SF. Her presentation is now online, and we also recorded a podcast with her talking about modern UI and how new technology gets accepted in society.
-
Sincerely,
Charles Humbe
Editor-in-chief, InfoQ
|
