The Future of Programming Systems – four thoughts | Hacker News Reader

The Future of Programming Systems – four thoughts

blog.waleson.com

5 pointsjtwaleson4 comments

I have a strong hunch that physicallizing the virtual (code) can unlock a new era in computing.

We see VR/xr as kind of on the downswing (to be honest), a troth of disillusionment phase. It's unclear its use.

Meanwhile personal computing has also peaked. Our digital interactions are defined and shaped by far off technofuedalists, and we have increasingly little say &power in computing. Open source exists, but it's a hard path to waken to, to get to where you can be constructing & shaping your computing world.

My hope & interest is that, if we can create representations of code and runtimes, if we can show how computing is actuating & happening, there's a hope we can re-open computing to a broader range of barefoot developers, tinkerers, and self-improvers, and start some new virtuous feedback cycles of computing being direct, engaging, empowering again. Which will make it interesting and hip.

We already have Minecraft and Satisfactory. The jump of having complex logic be entirely in game constructs to having it be data based, having hooks and endpoints to the outside world, isn't that wild. This idea of teams having views of their codebases, such as a big spatial wall shown in the beginning, is a start, reifies the virtual, is a physical twin of the digital (rather than a digital twin of the physical). And then we can show execution and state flow across pieces of the code, watch as a new value flows across space from its assignment to the declaration. Layers of abstraction can be black boxes, where we can see a macro view of data flow then peer into an object to see it's internal execution.

We have yet to imagine what living with our code might be, and convinced that there are chances for this esoteric art to someday become mainstream, interesting, and even cozy. This de-virtualification by creating representative spaces, I suspect, taps into how humans view and model the world in a way that can open how we engage with code and systems and runtimes.

I don’t see it as a single future, but the kind you r approach and tools a lot of problems will be addressed with. Some diagrams generate code. Interacting with some 3D models generate code. Some explanation generate code. And you can interact with more of those interfaces/methods and code to generate better code. But it won’t be the whole picture.

one line of code takes up exactly one line of text on screen. of which I have about 80, and I use two columns.

A visual language necessarily takes up more space.

I fail to see how a less efficient visual representation helps us look at more than a tiny subset of them.

That wasn't really the point of the article? The linked original blog post goes into that, but the point of this article was more about better representations and better abstractions in general, not about visual programming vs character based programming. For example, projectional editing can still be used to program in code, but what you're manipulating / storing is not the characters but the AST.

Reply on news.ycombinator.com