Wolfram Alpha has spawned a new symbolic programming language, Wolfram Language, with Stephen Wolfram revealing what he promises is the “most productive” way to create code. As with Wolfram Alpha search, which uses contextual awareness to answer questions, Wolfram Language is a “knowledge-based language”, its inventor claims, giving the language itself continually-curated awareness of the world it exists in and how things like interfaces are structured, allowing even those with no prior programming experience the opportunity to build complex apps.
“It’s a language where a vast amount of knowledge about how to do computations, and about the world, is built right into the language” Wolfram explains. “So, right within the language there are primatives for processing images, or laying out networks, or picking up stock prices, or creating interfaces, or solving optimization problems, or whatever.”
At its heart is natural language understanding (NLU), which can be used to specify real-world objects and concepts when writing software. That can include dates, locations, and units of measurement, as well as the standardized representations for “millions of real-world entities”.
For developers, each chunk of code is incrementally runnable, meaning sections can be trialled without demanding the whole app be completed. Local and cloud-based apps can be built, connecting to other apps, services, and devices through the Wolfram Cloud.
For instance, Wolfram Language could be used to create visualization apps that can crunch through big data without the programmer needing to first instill in the software an understanding of how that data set works and interrelates, or to program and control multiple Internet of Things devices and services as a common framework. Automatic face-blurring in a set of photos could be performed, or handwriting recognition across multiple styles.
There are more suggestions in Wolfram Language’s sample code gallery, spanning pure data crunching through more complex manipulation of things like Hipstamatic filters. Intel will also use Wolfram Language in its Edison SD-card-sized computer, which it first showed off at CES back in January.
It’s an ambitious system, certainly, and it works across Intel and ARM chips, on Linux, Mac, and Windows systems, and “soon” on iOS and Android, as well as select embedded OSes. Free versions will be offered through Raspberry Pi and certain casual cloud implementations, with more standard licensing structures for more professional deployments.