D155 - One source, as many binaries as the targets need. Single-binary was never the goal and is not a constraint to design around

Standing position, recorded because it was agreed in conversation and would otherwise be re-litigated by whoever next hits a target that will not fit the others.

The requirement has always been one source tree - a probe whose findings can be compared across platforms because the same checks ran. It is not one artifact. GEN=4 and GEN=5 already produce different binaries from this source, and host, payload and deck are three more.

That distinction is forced rather than chosen. EI_ABIVERSION sits in the ELF header and a loader reads it before executing anything - Kyty's test is literally e_ident[EI_ABIVERSION] == 2 and shadPS4 refuses with expected 0x00 is (0x2). A byte the loader inspects before the guest runs cannot be decided by the guest. Everything after load is runtime-detected and stays that way: 005-generation infers the generation from which graphics driver resolved, and the display takes whichever video-out pair exists, preferring the older when both appear (D111, D140).

So the rule is: runtime detection wherever the platform will answer at runtime, a build target only where it physically cannot. Nothing today is held back by this, and no split is being made on spec. What this decision buys is that the next thing which genuinely cannot be detected at runtime gets its own target immediately, instead of the source being contorted to keep one artifact working for two consoles.