You’ve felt that gravitational pull, haven’t you? That quiet, almost nostalgic urge to return to a niche programming language that just feels right in your hands. D language has always been that place. It was built to be a highly productive, feature-rich systems language—a breath of fresh air in a rigid landscape. But look at the DConf 2026 schedule in London, and you’ll feel a sudden, sharp disappointment.
The first day is heavily skewed toward LLM hype. It feels entirely disconnected from the actual struggles of the D community. A programming language cannot buy a ticket to the future simply by slapping an ‘AI’ sticker on its compiler. Chasing the generative AI dragon might win a few clicks, but it distracts from the existential crisis happening right under the language’s hood.
The tension is brutal. D was originally designed to give you everything: high-level productivity, rapid development, and safety nets like garbage collection. Yet, its most compelling modern use case today isn’t in high-level application development. It’s in WebAssembly (WASM) and bare-metal environments. And to play in that arena, you have to use ‘betterC’—which means explicitly stripping away the very high-level features that made you fall in love with D in the first place.
To survive in a bare-metal world, D is being asked to betray the exact design philosophy that brought it to life.
This isn’t just a D problem; it’s the universal struggle of legacy tech. How do you balance original design philosophies with shifting market demands? The market is moving toward constrained, low-level execution environments. Rust is dominating the post-Rust systems programming conversation. If D wants to be a viable alternative, it cannot rely on AI keynotes. It needs Phobos 3 to successfully execute a ‘betterC’ pivot.
As one developer rightly pointed out in the conference forums: if you want D to be nicer than C for WASM, you can’t just strip away the ‘nice things’ and call it a day. You need a default RAII vector class for betterC. You need a hashmap. You need the foundational building blocks that make constrained environments actually bearable to work in.
True innovation isn’t chasing the loudest hype cycle; it’s having the courage to amputate the features that have grown too heavy for the modern market.
D doesn’t need a keynote on how LLMs can write D code. It needs a ruthless architectural pivot toward betterC. The community is fighting for its identity and relevance. If D language can’t embrace the painful tradeoffs required to thrive in WASM and bare-metal spaces, it won’t be remembered as a powerful systems language. It will just be a nostalgic footnote buried under a pile of AI-generated slide decks.
FAQ
Q: Why is the AI focus at DConf 2026 a distraction?
A: AI integration doesn't solve D's core architectural bottlenecks. Developers working in WASM need foundational betterC tools like RAII vectors and hashmaps, not LLM wrappers that ignore the language's underlying memory constraints.
Q: What's the practical implication for D developers?
A: You have to embrace betterC, even if it means abandoning the garbage-collected, high-level features you love. The future of systems programming is constrained, and adapting to this low-level reality is mandatory for relevance.
Q: Can D actually compete with Rust in the systems space?
A: Only if Phobos 3 successfully pivots to a betterC-first philosophy. If D continues to rely on its garbage collector and high-level abstractions while paying lip service to bare-metal constraints, it will lose the post-Rust systems programming race entirely.