[Keynote] “Codename: Tunahead” aka Software Engineering Economics for Runtimes
summary
Argues that the runtime-economics that drove Lucasfilm to build SCUM in 1987 recur for WebAssembly: building a portable runtime plus toolchain beats rewriting for dozens of OS/arch combinations, but toolchain drift creates real costs — e.g. debug-enabled builds blow 4.5KB → ~50KB for hello world — and missing features force teams to maintain compilers. Uses the Bytecode Alliance ecosystem, Wally, Wasmtime and WasmEdge to extract a 120KB binary system-plugin that runs unmodified on both, showing a feasible $4.15B rebuild cost vs $8.8T value case and proposing a universal plugin API and verified runtimes for certification.
AI-written summary · about
chapters (14)
- 0:00Keynote introduction
- 1:26Maniac Mansion runtime origin
- 3:57Runtimes vs porting economics
- 5:44WebAssembly standardization challenges
- 7:10Compiler and library compatibility
- 9:20Open-source in embedded value
- 10:46Bytecode Alliance long-term support
- 11:47Wasi system interface duplication
- 13:53Linux-like system interface proposal
- 14:56System interface plugins feasibility
- 17:24Rebalancing costs with interfaces
- 18:07Mixed criticality sandboxes
- 19:53Tunahead codename origin
- 21:01Q&A concurrency certification
related talks
34mPanel Discussion: Browsers & Wasm
29mThe Return of Write Once, Run Anywhere
11mInterview of Mark Rendle
30mShrinking Software Attack Surface with WebAssembly & CNCF Wasmcloud
12mSponsored Keynote: The Next Epoch of Technology: Wasm & Actors
28mLet’s build with WebAssembly! by Tejas Kumar (Eng) | JSConfAM19
19mKeynote: Making JavaScript on WebAssembly Fast
53mWebAssembly