The AndPlus Software Development Blog

Consider the lowly umbrella: A mundane object, often cheaply made and inexpensively acquired, and with a singular habit of failing to do the job it was designed for. In anything but a light rain that falls straight down, an umbrella—even one of those big golf umbrellas—will keep very little of you dry. And if you’re sharing it with someone, forget it. As the Police sang many years ago, “It’s a big enough umbrella, but it’s always me that ends up getting wet.” It’s a wonder anyone uses the dadgum things at all.

As if we needed more evidence that machine learning is making its way out of the lab and into the hands of “regular” developers and their applications, along comes PyTorch, a Python open-source package developed at Facebook that enables neural network modeling, training, and testing, with a focus on deep learning and high performance.

It’s a strange irony: The more we try to make technology simpler, easier, more intuitive, and more convenient for end users, the more complex it becomes.

Consider the personal computer. The earliest PCs were simple by modern standards, with straightforward hardware architecture and minimally functional operating systems. But the user interfaces (C:\> prompt, anyone?) were opaque to anyone who wasn’t a computer engineer or hobbyist.

From the “Be Careful What You Wish For” Department

Can it really have been only a year or so ago that commentators, both in this blog and the mobile development community at large, were complaining about how hard it was to write cross-platform mobile apps, and wouldn’t it be nice if there were some way—any way!—to generate fully native apps for each mobile operating system from a single code base?

Ever since philosopher and mathematician René Descartes first set quill to paper to draw a line on his newfangled “Cartesian plane” in the 17th century, people have sought ever-cleverer ways to represent data in a pictorial format. The reasons are obvious: A graph, chart, gauge, or map can, at a glance, show important features and trends of a data set that you might miss by poring over tables of numbers. It’s the reason why Edward Tufte’s The Visual Display of Quantitative Information is still considered the bible of data visualization almost 40 years after its first release, and why firms large and small are demanding software “dashboards” showing the real-time health of their businesses.

We talk a lot in this blog about programming frameworks and how they help developers do their jobs in various languages. It seems at times that for any given programming language there is an endless list of frameworks available. This is great for us, because it gives us an endless stream of material for the blog.